#31 GPT 5.2, Security meltdown in React, Cursor visual editor, Figma becomes Photoshop killer

And we are live. We are Oh, it says on the screen. It says on the screen. You don't need me to tell you. Hello everybody. Well, uh, it takes a second to go live on Instagram. Man, Instagram live sucks because the delay is like 10 seconds. Live sucks. There you go. Uh there is something on your screen by the way. I see that. What? That there is something on the Yeah. What is that? Uh uh this this this thing. No, no. The other d the other way. This white thing on the other side. This white thing. Yeah. Oh, hang on, hang on, hang on. Cuz I think it's reversing it. So, is it is it the white thing that's down?

The white thing. So that is my monitor just to create a little bit of depth, you know. But yeah, I have that as well, but I you know, I can I can I can not do that if you think that's better, but I just like to Yeah, I think it is better without Without Yeah. Okay, fine. I have all my lights, you know, I don't have all my lights. So, let me also turn on my rocket lights. I don't want to ask. I don't know what to ask. I've got this there. There they are. Okay, cool. Hello everybody. Hello Instagram. Uh people on Instagram, please consider joining us on YouTube. You are live on YouTube.

You can see us in better quality with lower lag and you know everything much better than Instagram. I'm going to link it. Actually, there was something I wanted to say with regards to getting maybe it's maybe it was to subscribe. Maybe the transcript. No, no, no, no. There was something I want. I should have made a note. I should have made a note. Um, no. We'll get into all that in a bit. But the thing that I'm going to show you has nothing to do with the trans. Well, it does, but you you you'll see in a bit. We're just we're just piggyback offing my demo. Yeah. So, yeah. Um, but there was something there was something I was gonna say maybe to subscribe or do something.

I don't know. I don't know because people are really good because whenever we say if you're on YouTube right now, just hit the like button. Everyone does. And it's really cool because it really helps. It It does help. And we are closing on on uh 1,000 subscribers. Yeah, that's a milestone. That is a big milestone. And by the end of the year, we are hitting 10K. So, you want to be a part of 10K. 10K. Yeah. Well, we prefer you over on YouTube um just because we can chat to you and uh respond to your comments and things like that, whereas we we I don't even know whether you can. Oh, what's going on here?

Raising. Uh, yeah. Okay. Yeah. Um, I'm I'm trying to be a bit more healthy by standing. Yeah. Well, I told you about my new told you about my new my new chair that I've bought. Cost me like probably 600 something like that. Hang on. Hang on. 600. Okay. Goes all the way back like this. But can you stand up, mate? Why would I need to stand up? And when you can lay down. You can lay down. And and and as I say, it has like a a lumbard support, electronic lumbard support, which is now massaging my lumbar. So, how does it work? Like you have to charge your chair or like plug it in. It comes with a comes with a battery and a USBC.

Should I do the rest of the show like this? Yeah, sure. I love how you could just see my head. Well, I could sit down as well. Um, and what else has it got? Well, that's kind of it. but it has a foot stool and they're really leaning into this um you know whatever it is this this feature cuz it comes what's on the way is a blanket and an eye mask. They were stretch goals of the kickstarter that I supported and it come it already came with a foot stool which I have to me just over here and it's just nice just to kick back and so you have to char you have to keep your chair charged.

I mean it can work without it. It's just if you want to, you know, cuz the lumbard, you can electronically control the lumbar thing. What lumbar, I don't know what it's bloody called. Once it's set, it's set. I don't need to do it. And it's just that massage feature really that I care about. Um disappear if the battery runs out. Yeah, I'm just being uh Yeah. Uh anyway, I I think they should have done it with like char charging it with the the wheels or something, you know? Six of those as I'm as I'm rolling around the office, you know, it's charging up. Yeah. Yeah. Well, I'm um I'm fairly I'm fairly happy because I figured my my one was slowly I mean it might have even happened on the show for the first time actually, but my old one it was just slowly going down.

I'd be like bloody hell and I'd lift it back up again and I'd be right and then it would slowly go down again. Um, so I got rid of that and I was like, right, if I'm gonna buy a chair, I'm gonna spend a bunch of money on it and I'm gonna, you know, buy one that's going to last because it did. That one lasted lasted a while. It was only a 5050 pound gaming chair. You're you're sitting in your chair longer probably than you're spending time in your bed standing. Yeah. Yeah. And also I bought bought a a nice arm now for this camera. So it's slightly closer to my monitor and it takes up less less desk space and it's like it's just clean, you know, it's clean.

Anyway, anyway, shall I shall I reveal the surprise that I have or should we wait a little bit longer? Uh we we have some people here. Okay. So tell your friends people that we have a surprise. I don't know what the surprise is. Sam has a surprise for us and well actually and it will lean into and I think this is what I wanted to say actually when I was like oh I was going to say something right we are encroaching on a thousand subscribers on the YouTube and obviously that's a big milestone but we we were just discussing before the show we want to bring you more we want to do more exciting things and if you are if you have and we don't mind I don't mind hearing stuff that we could do better.

It could that we just don't we stop talking about this at the beginning and we actually just get on with the show. If you have a suggestion or an idea, if you want us to do something, we're we're really looking to like obviously up the ante now for the next thousand subscribers and bring more maybe more interactive elements, maybe more um just something. I mean we we what happens on YouTube is that we unlock memberships and obviously running a community is another you know in of itself and whether we can actually do that but if you have ideas if you've seen other creators do something really cool maybe it's merch maybe it's I don't know tell us and we will listen to it like like help us out a little bit equally if you want us to not do something if you if we do something that really really annoys you more than happy to hear it criticism is amazing If it's constructed, talking about AI.

Yeah. Yeah. Yeah. Yeah. Just stop talking about AI. Just stick with the no code crap, guys. Stay with Web Flow and Framer. Gosh, all this AI Um, yeah, more than happy to hear constructive criticism. We We don't take it offensively. When people just say something and it just sounds like they're being angry and upset about stuff without really giving us any guidance. That's that's what's difficult to even respond to in a positive way to turn it around into something constructive. But if you have something constructive, more than happy. Yeah. Um, so the surprise surprise. I'm I'm I don't even know how this is going to turn out. And I have a feeling this is um this might not work the first time, but but it's okay.

So because I know the change that we're going to make. In fact, I'm going to do something real quick so that we can quickly um make the change necessary change. Uh are you listening for test events on your NAN? This is nothing to do with the surprise by the way. Just are you listening for test events on your I I clicked on listening and nothing. Okay. Yeah, that's fine. That's fine. That's all I need to know. Right. So, to let you in on a little secret, guys, we have whenever we go live, um, I put in all the timestamps for all of the topics that we cover in the description of the video. Once I've done that, I've set up this N workflow that when I click this, it goes to the video, picks up all the tests, you know, all of the topics that we talk about, gets the transcription of what we talk about, and then I run it through an AI to generate I'll I'm quite happy to show you the prompt.

Um, a social media manager to generate LinkedIn posts for all of the talking points, which I then post on my LinkedIn. Keeps the conversation going. I can link to the the clips that we do and various things like that. Realistically, we can do anything at this point. And this is where this is Kabaza. This is why I was thinking about this could lead into some ideas that we have because realistically, what could I do here? I could say write some blog posts. I could say um well, yeah, a mailing list or something like this. We could write an email about the new show. We could whatever. And and and I feel like this is uh and let me know if you disagree.

I I feel like this is quite moral because it's not AI slop that's just lazily generated. This is based on things we've actually said and actual talking points. It's just the same as if we got a personal assistant to sit in on a thing and just write up some blog posts, which before AI you would you would often do. So with that, I was like, well, how can we take this to the next level? What can we do? Well, I hope you don't mind, Kabaza, but I've cloned your voice. Oh my god. I've cloned and because I have all of the necessarily um ingredients to be able to clone your voice and I've signed up to 11 Labs and I figured what what if we release a roundup called Command AI shortcut and it's a maybe it's a 10-minute maybe it's a three I don't know but it's a a roundup podcast that's just a short little condensation of what we've spoken about.

And if we go into here, so I've got a um I've got a thing here using the transcript cond transcript below. So the transcript that I've already got consolidate into a a short thing. I've told it we're using 11 Labs. I've told it that you can do laughing express. I don't know whether you you guys are aware of 11 ladders, but you can add all of these annotations in that actually add sound effects and add uh tonality and we've actually spoken about this. So, it's really cool. You can even jump in as well. I can even cut you off all this sort of stuff. So, and I've always told it to end with what we end with.

Right. So, I've got this So, I've got this thing and um it's all triggered off of uh this whole workflow. So, I've got the transcript and whatever. And this is why I said run the edit end thing because I'm about to send you this and you can deal with it offline, right? And then I sent it to 11 Labs and generate a audio version of the condensed version of the command AI shortcut. So, I was just going to run it now. It's going to do its thing. Uh, it's probably going to get last week's episode. If it doesn't, it might actually break because it might get this week's episode. I don't know. Uh, yeah. So, I think it's got this week's episode, which is ah, that's really annoying.

Let's um Can you see that it's got like 31? Yes. So, let's let's make a quick change. Um code in JavaScript. Could we RSS? So, basically I get all of the episodes and I filter out the first one. So, let's do um let's do this pod cast because I think it will be the second item in the array. So, we could potentially just do that. And let's just run that to see if it gets uh RSS feed. Oh, that's annoying, isn't it? Yeah. So, it's got the You got time. Okay. Well, thanks. Thanks for the time. Um, I'm trying to work out how Yeah, we can do whatever we want. How How is this feeling for you right now?

How what are you thinking? Is this something I don't I don't know. I'm I'm scared of like my This is the first time. No, this is actually not the first time my voice being cloned, but um Yeah, I don't know. I I also gave my voice to Apple. This was quite a long time ago. Mhm. Uh they had this thing to clone your voice for a time when you might need it later on. Oh yes, I remember. Yeah. Yeah, I remember. So, uh gosh, I'm actually very excited to to hear that. Probably with better accent and much, you know, much better grammar for sure. Yeah, but that you can't you can't give it in the prompt.

You Yeah, it's taking time. Use like B1 level grammar. Uh, how do I do this? Okay, so can you just change um I've hardcoded it. So maybe ah okay. So execute workflow because it's got all of these, right? And then I take all of them. Maybe it needs we have released a lot. There it is. So it's there. Okay. And then as you fix that for people who might be watching and not knowing what what this software is. This is N8N. This is for automation. We use this to automate pretty much anything. Uh my favorite automations are mostly with Telegram mixed with Telegram. So you can you can use Telegram to send a message or to trigger something and then a chain of things can happen.

And what is like super duper cool about it nowadays about automation is that you can have something called AI nodes. So just imagine just an AI in the middle of your automation making decisions for you or like you know producing something for you. Previously without AI you had to when you would automate something you would have to match things exactly and specifically and say like if the user says exactly this you would have to really match things. But now with AI, you can take something from the user whatever and and ask AI with a prompt to to decide what to do based on, you know, analyzing that thing. So it is yeah much you much more flexible in a way.

You can do much uh cooler stuff with it. And in this case, Sam is trying to automate generating posts for LinkedIn based on the conversations that we have here. So basically taking these conversations and put them in chunks, smaller chunks um for LinkedIn uh which is which is really cool. And for people who are interested in this by the way this is like a whole thing in itself like companies actually pay you to do this kind of things. Um you can imagine like companies do a bunch of things manually right and there are people in companies you know going to work day in day out to just manually do some certain tasks again and again and again and uh of course like it would be uh something that companies would be interested in to pay for if you could uh automate their their stuff.

So yeah, so you can see it's running. So on the back end, you should have got you should have got a post event to your NAM workflow, but again, you guys don't have to worry about that. That's just offline. So what it's doing right now, I've sent all of that into the AI agent, which is now generating that script of a shortened version of last week's episode. So the output of that will be what I've told it to be a script. So we can even see it come through. So there we go. Broadcast start clears throat. So again, you've got all these things. So clear's throat. I think it's going to break. Oh no. Let's have a look because I I'll give you the insight.

So now it's sending this through to 11 Labs to to take that script and generate an audio version of it. Now the reason why I thought it would break is because taking our text and creating a voice out of it based on my voice. Yeah. So, I've cloned Cabaza's voice here, if you can see that. I've cloned my voice, and I I wasn't too pleased with the thing, but you get one credit to create a pro voice. And I've given it like two hours of my voice. And this this is the And but when I hover over this, it doesn't say version three. It only says version 2.5 on all the training. So, now that I've sent it to this, I thought it would say you cannot do this sort of stuff with version uh 2.5 of the provoice.

So anyway, point is it seems to be working. It hasn't thrown an immediate error. Um so can you show us the N8 again? Please show this. Yeah. So right now, okay, it is still like there we go. So it's done. So actually there's something happened here. So um we need to figure that out. Um so it's done. So let's do this. Let's download this file. And let's see if we can open it up in the browser so we can all hear it and be incredibly embarrassed of the potential racism that could come from over. Yeah, we last time we did it, it was pretty racist uh the AI, but you know. Okay, I can already hear it.

It sounds But but can you see this? Yeah. Yeah. Right. We're going live. We're in. Good evening, good morning, good night. The shortest live ever if Cloudflare has anything to say about it. Did you see the chaos this morning? The internet just vanished. Web flow streamyard. It's pretty good. Like you're pretty good. Scary how lost we are when one server goes down. But hey, at least we made it. But speaking of panic, Samuel, there's a massive dread right now over at OpenAI. They are freaking This is three and a half minutes. Oh, absolutely panic station. Does it sound like us? Sam Alman reportedly called a Yeah, guys. Does it sound like us? Is that because this stuff gets better, right?

Is that possible for a command III shortcuts or something where you can consume the news in three and a half minutes? It will obviously be podcast, but I've no doubt we could again using nan we could feed this into um you know uh what's the video one called? Um I forgot the name of it. An andia. What was it? Yeah. And yeah, we could feed it into something like that. We could write a blog post on this whenever. But the audio version could we synthesia synthesia. Yeah. Yeah. Yeah. But yeah. Um I'm not sure that was worth my one credit of training. I actually thought the older I actually did run this through the voices before and I would say the one that it's not been trained well the one that's been trained the least sounds more like me to be honest.

But there I think there's a tonality in what we're saying that you can kind of work out who's who. It could be something we do. It could be something we do. Yeah. This this is so my impression uh it is better than what I thought. Right. Okay. But I feel like uh there is this long tale that we are keep ignoring or we don't see it yet. And I feel like this is very similar to self-driving cars example where we could do self-driving cars a decade ago like the demos but like the actual long you know the long tail of the small details makes it or it has made it very difficult to have like actual self-driving even today.

Um, and this is the same I I I feel like this is the same with images, with video. Even with images, I feel like we are not 100% there yet, even though we are super close. But with video and voice, it's just much more difficult to have that 100% authenticity. But I don't know. Um, I mean, it it is it is better than I thought. This I mean, it raises the question of like what what was I So, okay. So, I saw a post on LinkedIn and one guy was like, uh, all of the applications that they're receiving for some grant are being AI generated, right? And my response to that was so you can't avoid you you can't just say to people, no, you can't use AI, right?

People are going to use AI. So, that it's here. The the Pandora's box has been open, so we have to deal with it. My response to it was we need to understand the reason why we are running these application processes. What is it we're trying to get out of these people? What is it we we land why did we land on the form as a way of determining or or answering those questions? Where I'm going with this is what are we trying to achieve? Like we're not this is our art. This is this our live stream is our art. The thing we're putting our ideas and stuff like that is we're not we're not looking to replace what we're already doing.

We're just we're just using it as a platform to broaden our reach and provide an extra service. Like I you know you you probably got your own opinions on this stuff but like my posts on LinkedIn I'm again I would get an assistant to do that. I I'm looking for good, but I'm not looking for absolute perfection on every little thing that we do. We just physically don't have the time. I want to improve our edits, but I physically do not have the time to just make them all every everything that we do absolutely amazing. So my question is, is this good enough to serve the purpose of a three and a half minute roundup of what we spoke about or is it truly detrimental to our brand that we have something that that goes out that you know whatever?

Let's say for example, yeah, I see what you you see what I'm saying? Like what are you looking for from something like this and why should it meet such high standards? Uh, let me surface one of the comments. Um, here. That was horrible. Did it Did it work better with just your voice, Sam? So So what do you mean by that? Are you saying when I just ran it through and it was just me? Cuz it sounded better on the because you can you can clone your voice with two options. You can clone it as a pro and you only get one of those and I need to upgrade to like 300 because it does all these pre-train.

took like six. It took quite a few hours to build that thing based on two hours of recording. And there's another option which is instant and you upload like 10 minutes of voice and it does it in like less than five minutes. The one that sounded was less than five minutes sounded better than that one. So I'm not sure what your question means there, but like it doesn't change whether it's just me or whether it's a dialogue thing. It uses the same voice. Okay. So for me to answer your question um and also to uh talk about the comment I I think your voice was pretty good um mine was not really as close and to answer like the main question is this going to harm our brand I would say yes I would say I yes there are some people that might you know listen to this and not notice that it's AI or they they might and not care.

But at the end of the day, I I feel like this will do more harm than good. Um especially like right now, I don't see people being very kind to AI. Uh, I watched this video about Microsoft going allin on AI and users absolutely hating it and it is just about like people don't really uh like this all going in AI thing. Um, even though in some cases it is it is useful and I I do get the appeal three and a half minutes of you know condensed everything. It's much much easier to do it with AI. Like if we want to do the same thing by editing the the show down, h that would take ages, right?

It would take an editor like ages to to condense everything. And if they could if they even could because it's pretty difficult to take out so much context. But yeah, um I think it is super cool. I personally wouldn't want to put it out there because to me it feels still a bit um too unpolished, too AI sounding and to I would say if somebody doesn't have time um to to watch, you know, like the whole show, a a text version of that would be probably maybe better like have something like maybe we can use AI try, you know, to to get what we talked about and generate a much much shorter text version and just put that out there.

So, if somebody wants to just know, okay, what is up in one minute, something like Morning Brew, I don't know if you read Morning Brew or not, I I I check them out on Instagram. I go to my Instagram, I see Morning Brew, and I know what is up. And it's very condensed, very short. So, um, that take on it. Yeah. Yeah, we we we'll take it offline because I think like you know the question because I've got an app that on my iPad that just reads out the page in the voices that they provide to me. A lot of people do digest things in audio format. Again, do the the pros outweigh the cons?

So, we we'll move on, but I will show you the a version where it's the quick version of my voice. So, if I go here, I think it sounds better. you let me know. Right, we're going live. Good evening, good morning, good afternoon. It has been a chaotic start, hasn't it? Cloudflare literally went down this morning. I couldn't even Let's just quickly go back to this one and the Cloudflare one. Right, we're going live. We're in Sam Alman reportedly called a code red. What do you think the video? Because the the the second one the second one definitely has more depth to it. This is the quick one and then I think yours is exactly the same two model which is open source and on par with there there are moments where you can kind of hear where it's going with it because it's a specialist model for scientists and mathematicians.

So yeah I I don't know my my issue is not with the voice exactly because you you can tune voice to to make it match better. Were you Were you hearing that by the way? Because I don't didn't see Yeah. Yeah. Yeah, you heard it. Okay, cuz we weren't in this view. That's all. Cool. I did hear it. I don't know about the audience, but I did hear it. Um Um Anyway, to me, it's about the like the the tonality, the the way we say things and the I don't know. I I feel like there is something about that, something human about that, that connection. And that is what matters to me. I I I mean, I mentioned this before.

I don't want to watch AI generated content and if I do happen to see some it is generated by a human for a purpose like trying to communicate typically a meme or something like this. To me, this is a bit a step too far in the direction of completely taking me out of the equation because I know it is based on what we talked about, but it could be also, you know, given it anything, right? We could give it any That's my friend. Um I don't know. There there's an element of me that sort of argue like wants to say like get off your high horse a little bit and like you know you have a team working for you.

They represent you know they do things for you like I don't know I don't think we're all that if if if you were to sort of shadow a um uh yes I'm split on whether it was better or not. I mean you know then it's just fun isn't it? But um glad you heard it. But like you know if you were to shadow like a CEO of a of a company, right? The amount of decisions that get made that that you know if you you get ghost writers, right? You get all these PR people who literally they just sign stuff off and then all of a sudden they've said this thing like there is so much deception that happens whether it's AI or not.

There's so much deception that happens. So I don't know. I can hear your cat power. It's really really peaceful. Let's move on anyway because I think but it was just a fun experiment. Just one last thing about that. Go ahead. Um and I and I think this is for anybody who wants to do this maybe for their clients. There is an opportunity for that. I have a client who pays pretty good, but they are super super duper busy. Like they are extremely busy and they they've been telling me this for quite some time that they prefer having something on LinkedIn than nothing. I don't know. I'm also like split on where I stand with this because sometimes maybe not having anything is better than having something really bad.

But they want to generate content uh with AI and just put it out there video, audio, text, everything. So yeah, um once they have time, I will talk to them about maybe maybe even we do a collaboration with them, you know, generate stuff for them and put it out there. But there there are people who really don't have the time and they see the value in just putting out something that is resemblance of what they think. Um so there is so it's out there but at the end of the day um we have only this much time and I don't know if people will actually consume it. Yeah. Cool. Let's go on the side of the week.

Cheers for humor and me. That was a really good fun uh and I'll figure out the N thing, but it was a good like experiment that I was thinking about like how can we bring more how can we bring people content that's valuable because end day it needs to be valuable to you guys right and if you think that sometimes you would just like a little whatever and you are the type of person that consumes content audibly because obviously there's not video we could add text to this. This could be a blog post that we add on a website that doesn't does or doesn't exist yet. you know, the the text is there, but you could press play and all of a sudden you're you're listening to the audio of it.

So, yeah, let us know if that would be a valuable thing. But yeah, good fun. Uh yeah, I I I could see the value in that actually. I could see the value in that's an wait that that's actually a good idea. Imagine on go like um you know the web flow podcast. I've been on the web flow podcast and what they do they talk to you and then um they edit it and they put the the audio version out but also they have a text version. It's the transcripted version. So for us we could have like the text but also an audiogenerated version of that in the same page. You know how I like I sometimes use just what Google Chrome has to to read it out loud for me because I I want to read an article.

I use Yeah. Yeah. Exactly. Right. I can't read it because I I can read it because I'm like on my bike or, you know, in a car or something. I want to listen to it. So I that that's how I do it. Now imagine it's the same thing but with our voice. It just two different voices, you know. Would you would you feel better on one of their own voices, you know, where we could select two random a boy and a girl or something like that, you know, like they're probably better recorded and probably have a better result, but but that's the point. It's it's about what where the value is. Let's establish what the value is rather than getting on the high horse about the use of lack or lack thereof of AI.

It's not really about the AI. It's about giving people a format, a new format that that might be more suitable for them. Um, yeah, as an alternative, I do like it. Yeah. Yeah. Uh, oh, yeah, definitely an alternative. Uh, cool. Nice. Cool. All right. Uh, it was a long pre-intro thing. Uh, but it was worth it. Worth it. It was It was fun to hear the voices. Now, let's go. Uh, should should I start with the intro? Yeah, I'm going to chill out while you do that. Okay. Okay, I'm here. Yeah. Okay, so this week your camera is focused on the Okay, now it's focused on you. Uh, cool. Cool. Okay, let me see who do we have.

So, okay, we have like 52 people and now we are officially starting with quite a lot of things uh this week. So, this week OpenAI released GPT 5.2 too. Uh, React got hacked really bad. Anthropic dedic uh donated the MCP protocol. Uh, cursor released a visual editor which might put Web Flow and all other design tools in trouble. And we will talk about some drama, a $50,000 landing page. Uh, all of this in today's episode. This is Command AI. I'm Karza. And I am Samuel. And this is a week in AI design and death. Cool. Yeah, it it looked like uh Vsauce video. And we Hey there, Vsauce here. Where? Down here. Okay. So, cool.

Let's start with some drama. Um All right. So yeah, we we have some Twitter drama. So I'll Okay, so I'll take that from here. So $50,000 for a single landing page. Let's talk about this Twitter drama. And I can't add my screen for some reason. Wait, is it because I'm sharing it? Ah, I need to remove yours. Yeah. Okay. Okay. So, can you charge $50,000 for a single landing page? And this is the drama we are going to talk about here. Marcel uh said, "Earlier this year, I was charging $1,200 per landing page. Now, we have we have single landing page clients for 50k. $50,000 for one single landing page. Uh he tweeted that and obviously people reacted to this.

Uh and then he was like, "You don't get it, guys. A company that raised millions investing 50k in a worldclass landing page is normal." But yeah, people reacted to that as well, including Brett. I don't know uh if you have heard of Brett or not. you probably have. Uh he retweeted a screenshot saying that I'll pay you 50k if you show me proof that you actually got 50k for a landing page. And I think this is very important for a landing page. We're not talking about things around it, just a landing page. And I don't know how you feel about that but uh I this might have happened but I feel like charging 50k for a landing page only is a stretch by a long shot and I have worked with companies who raised millions like literally millions with uh enterprise clients who don't mind paying like huge amounts but 50k just for landing page.

That's the point. It's It's just a bit unbelievable. Anyway, so Brett said this and you would think that Marcel would provide um proof and Brett would, you know, pay him or something, but it seemed like that the drama continued because Brett uh blocked Marcel after Marcel tried to show him some proof or something like that. So, so they went into the DMs and uh essentially Marcel is asking Brett, "Hey, what type of proof do you need to believe me?" And he's like, "Uh, I can send you the Figma, the payment invoice, company name, and most importantly, sufficient evidence that the payment was made for a single landing page." This is what Brett is asking. And I think this is pretty, you know, reasonable to ask like is it just for the landing page.

Anyway, so apparently uh they did a little bit of back and forth and at the end Marcel is saying, "Okay, I can definitely send a payment statement and if uh and Figma and clients company not launched yet, uh I'm assuming you weren't you won't share this with the public." So yeah, there there has been some drama around it and yeah, minutes after Brett blocked Marcel. Both of these guys are really pro- rage baiters. They are really good at rage baiting and getting like hundreds and thousands of views just you know from um saying like out yeah think things like this. Anyway, so the question is, can you actually charge 50k for a landing page? And I don't also just mean one time maybe a specific case with a huge rush fee because you have to do it really fast.

No, I'm talking about like regularly, semi-regularly with big clients. Do they pay 50k for a website? So that's a different question. Yes, absolutely. clients pay far more than 50k for a single website. Myself, the biggest uh not invoice but the uh what's it called? Proposal. The biggest proposal that I wrote was for almost 500k like half a million. But it wasn't for a landing page. It was for the whole thing. It was like backend, front end, a website. It was like the app itself. It was a bunch of work that you know would require such a budget. Um my project by the way didn't move forward not because of the money but because of other things.

But yeah the clients are willing to pay this amount of money but 50k for a landing page. Uh what do you think? Well it kind of goes into what you just said about like everything right. What what is what what is Brett uh arguing or what what does he suspect is was was 50k uh sorry what was purchased for 50k because of course like is he saying so does he say anything like so the Twitter like if you look at like the replies and such uh people were mentioning that the the work that Marcel has is not, you know, on the level that you would charge 50k for. I I don't care about that. I don't want to I don't even care about the result.

I don't like what I'm interested in is well, first of all, I just need that I'm curious to know what what what Brett was, you know. So, so Brett uh is suspicious of this for a landing page. That's the point because I don't I don't think Brett doubts that 50K entered this guy's account. I don't think I don't think he doubts that. He's just he's and it and and it comes down to semantics. Well, actually, no. They paid you 13K for a landing page, but 10 10 of that was rush fees or blah blah blah was you know, I think I think what Brett is Yeah. Does he say or or mention anything about the does he mention anything like that?

Like actually it wasn't 50K for a landing page. It was 25K for a landing page but overages or something like that. Yeah, this is just the the tweet he okay he sent and there is not much around it but again like people are not questioning the 50k for a client which is not a huge number by the way like yes it is big but like for high-end clients this is kind of like the lower end actually. Mhm. What is in question is one landing page just the landing page costing this much. Yeah, we're talking about branding, right? There could be branding assets, um, development, a lot of things going into it. And I still don't know if this is like for design only or design and development.

But even if it's design and development, 50K is still very very high. like for one single page um for one single page like 20 30k would be normal design and development for something with I don't know a bunch of like crazy animations and something like that but 50k is a bit much so there's a couple things we need to get clear right and this is the classic trap a lot of us fall into myself probably included right those especially when you start your own company you quickly. The there's two things there's two things people often charge for. One is the this the the technical expertise, right? The the ability to co like ability to code a good learning page or the qu the the the design expertise or whatever and you think of it like that and you think well if it's good design or if it's good development then you can charge a lot for it.

The other aspect is which we're often once you start getting into running your own company and this and that you you start to charge for the the product itself like I don't need to know I don't I don't want to know if you think that's a good or bad designed website or if the code is good or bad. Um it's it it's simply that someone wants a landing page and you look at what the val it's called value based pricing right is what value what what is on the line of the success of this landing page what really is in on the line and you're told that if you're entering conversations with your clients you're told to really establish where are the wins what what would be this what are the success metrics and what is on the line here if this actually goes really really well because you could because Nike could be launching a new shoe that um is sponsored by one of the biggest sports stars in the world and they just need a landing page and if they get it right that what's on the line is potentially hundreds of millions of pounds or dollars or whatever you want to call it, right?

So you say well uh you know we're going to charge you 10% of that which is already 100k. So, like, yeah, when when the stakes are so high, that's where you start to get in the realm of actually it's a steal paying 50 grand because actually that's that's 5% of uh of the potential gains of this this product. Um, yes, but if that is guaranteed and there is no guarantee, that's why and that's why you don't exactly that's why you put 5%. And that's what we said. We can't guarantee you 100 million, but you know, that's what Yeah. And and just the the last thing I'll say is because I've worked on a 50k landing page, right?

The last thing I'll say is it's not just the design. It's not just the development. It's also the copyright. It could be the assets. I've worked on a landing page where we needed 3D assets made. The the the development work alone was I think between 12 and 15K, right? Not a lot. But that was through an agency and so they're going to take a 20% cut on top of that. So probably you're talking about 20 grand alone just just for the we also design we also designed the thing. We also wrote the copy for the thing. We also had to figure out the hosting and the whatever like if you were to go strictly on the technical expertise and charge for that you know there's so many there's so much that could go into it.

There could there might even be photography that might have been created. There might have been previsualisualizations. I'm looking at animations on your screen right now. There's a font. Maybe there was font design for it. You know, there is so much that could go into a landing page. It's worth establishing. Well, what were you asked to do? What what what constitutes as a landing page? Well, it was font design. It was it was a branding system. It was uh they could have had nothing. So, yeah, that's all I'll say about that. That's exactly that is the point. Uh 50k for a product design and landing page that's actually extremely cheap, right? Uh like this this level that we are talking about like this level that we see.

So I see like like good like good work. This is like really nice work. really really take taste out of it though because something could be gold to someone and and charcoal to someone else, right? So I'm trying to I'm trying to remove taste from this whole discussion. But I'm adding taste to it. Yeah. Yeah. Sure. Sure. But you can't judge you can't say something is is or isn't worth 50k based on taste. Like you got to go with imperative emperic. Yeah, I know what you're trying like what what you can measure. But even taste is partially like the design is also partially you could measure. But nonetheless like what when we are looking at a landing page for something like this there are asset design.

So if that is included uh you you could you could charge more but also the assets are they like 2D or 3D because recently we did some work and I um unfortunately I can't share it yet. It's not finished, but we did like 3D asset design for the landing page and we did the landing page design and it was a huge amount of work and it wasn't 50K, but it was it was like really really like it's one of the work that we are really proud of and I just know how much work went to in went into it. That's why I'm mentioning like it can be like 50k when you when you consider everything in it, but when you just consider the landing page and nothing in it, that is maybe a bit of a stretch.

I don't know. But nonetheless, u the work here is pretty nice. I wanted to show one other example of a landing page that didn't cost 50K, it cost 90K. So, you probably have heard of this. Julie uh Alex uh here shared on LinkedIn quite a while ago that they charged 90,000 that's nine with a zero, not 19 90k for Huly Labs landing page. And then he mentioned that they did branding, design, development and motion. And we can quickly look into it because this 90K project was actually like the one of like the thing one of the biggest projects that we were looking at when we did that uh project website of uh that project for my client where we did like 3D and design and everything for them.

We wanted to do something like on this caliber. So if we look at the website, you see like there are some interactions. You see there are some WebGL stuff like some like nice motion. Uh if you look at the product in a just just looking at it, it might look like oh this is just a screenshot but I uh assume it it there there are a lot of work that went into this. They there are like some lighting effects on this not just one layer like multiple layers of lighting. They might have recreated this inside of Blender because I know that they use Blender for a bunch of other things like all of these. They have motion, they have interactions and this is like made in Blender.

Same here. Like they have made like a whole section in you know like the assets in Blender all with this brand. Same here. Uh quite a lot went into this. Like you look at this where that 90 grand went right. Exactly. Uh when you see this level of details, you see they they didn't just added 8th of March here, they they went in and they thought like how can we make this look absolutely beautiful and they made it they made it really beautiful or like this sync in real time. You see like the amount of detail going into this is insane. So 90k it's it's like very understandable. very understandable. And just wait for me to show you.

It's not actually just one landing page. It has like this pricing page with it as well. Like look at this. This is the craziest um pricing page I've ever seen even one year after the launch. And they, by the way, for if you're interested, they did this with After Effects. So this is a video. This is essentially a video. They I I believe people recreated this with Drive, but at the time they did this, they did it with After Effects. And how many So if content updates, then they're Uh, no, no, the the the the text is text, but the background Well, if you added another bullet point, you'd the video. The the the card would get longer.

Yeah. Yeah. Yeah. I understand. But I mean but but but since this is developed inside of After Effects they they can easily translations what if you translated that to German. Yeah I don't know that's that's a good idea. That's a good question. I don't know how it is. Um if I um just just one thing to add though and we've already kind of touched on it anyway is like you know it's easy for the design community to be so they're so visual they often forget the the stuff that's unseen first and foremost being the strategic application of this like uh you know you can charge anywhere between 10 and 50 to 100k just for the strategic element before you've even lifted a pen or piece of pen and paper or touched your keyboard, you're charging upward, you know, up to 100 grand just to figure out what the hell we're doing and and how we're going to how we're going to execute on this in an effective way.

Yeah, see it's broken. So, yeah. And then what about mobile as well? Surely mobile like a tiny little iPhone, you know, um SE or whatever or mini will will make the text longer and it will break it. But um but anyway, you know, it's easy for us to comment on things for it to look pretty and we go through trends, right? We'll look at something again from a visual standpoint. We things that things look that look trendy look expensive, but it's just because they're trendy and they'll fall out of trend in next year and then it'll be something else and then it won't be deemed this this sort of aesthetic won't be deemed as as premium.

Um point is there's so much that goes into unseen stuff and again from a business perspective you know 50 50 grand adds up very very very quickly 90 grand even you know and and this is just one last thing on profits like difference between profits and revenue why people share their revenue I don't know because yes you might have charged 50 grand for a website but you know what in hiring people it might have cost you 45 so actually the profits are minimal whereas this could They could have charged 90 for this and actually that only cost them a 10 10 grand to make. So their profits are 80 grand. So like people just forget about the the business and the strategic side to stuff.

We've got a comment here. Couldn't even imagine the meeting you need to explain to get the green light it would take. Well, Bacharus Bakerus um you know to I think this is laughable to executive. This conversation is laughable to executives at Nike, at Adidas, at Apple. Like, they're just like, this is just the way it is. Like, we are we're playing around in the white water. If you're a surfer, you know what that means. Like, we're we're messing around in the white water when we talk about 50 grand. It's like, what? 50 grand? That's a bargain. Like, the salaries alone costs like 45 like Exactly. Yeah. Yeah. Like, yeah, I agree. But this is why I don't like uh valuebased pricing.

I I believe it is very deceiving and uh it is harmful to people just starting. You could sell a landing page for 50k, but that is not something that you can do often and this is not something that many people can do. What many people can do, what something that you know anybody who's good at their craft can do is price their craft based on the amount of detail that goes into it and the amount of work that goes into it. And this is I think what happens more often. Uh I don't think like with like big companies or startups anyone is willing or like most like but by far majority are not willing to pay value based pricing.

They don't they they probably don't want to pay you 50k for just a landing page because I don't know you happen to have like good taste in the past. So what do you guarantee? Why why should they pay you 50k when they can pay 10k for the same landing page? So you pay for different things like again like maybe some people here are focused on you're paying for the design. you're not you might be paying for um this person has a record of of landing pages that are performing really really well from a data perspective. It doesn't matter what it looks like. It might be all in the copy. So we're just we're just Yeah, but that so locked into the visuals.

Sorry. The copy is different. When I'm talking about like the landing page, I'm talking about typically when we say a landing page, we mean like just the design of it. No, no, that I don't I and this is why this was in the tweet. That's that's what I would get from it. That they they said we got paid 40 grand just to design this landing page. No, no, they that that's what I'm saying like they didn't say just the design, but that is what Brett's point here also was. So that goes back to my point about semantics. That's why I was trying to establish Brett talking about it. is is is he locked into a particular aspect where he can he can dispute no it wasn't 40 grand because 10 grand of that was copy it's like well the original person never said it was just design but yeah um semantics are a funny thing right but yeah it goes it goes back to the product of a landing page like don't just think it's just all the stuff you can see it might have been actually it cost 50k because they needed to fly them in to a different country for um the preliminary meetings, you know, the hotels alone could have cost five grand.

Yeah. It's all this sort of stuff that people just again I say it again, we're just playing in the white water when we're talking about like we have no idea what these these companies are willing to pay. Um, don't we like, okay, so just in my head to wrap it up, you can charge 50K for a landing page, but that includes a bunch of things, not just, you know, yeah, making a bento box. It includes it might include also web brand. It might include motion and a bunch of things. And my point here why I don't like valuebased pricing is if you price it based on those things that you are delivering to the client, it is much easier to see why it takes 50 grand.

And that is why you are also partially saying uh it is laughable to these big companies because 50 grand. Of course it costs 50 grand because you know the the the people that they are paying cost far more than that. People who are you know working at those companies and 50k is a small portion of that. And when you see the design, the copywriting, the the visuals, the branding, the the versioning like you would probably do many versions and the time and the constraints that you have. It will end up requiring so many people and so many hours that it is like very clear that it costs like in this case 90k because they had to have somebody to do the visuals, somebody to do the development, somebody to do the uh 3D like at least I could see like four or five people included in this project for many months.

So yeah, of course it cost 90k. H well the the debate is uh is still going to rage on like people are just going to think that's nothing for 50k. Some people are going to think that is the I can't possibly ever imagine charging more than 5k for a landing page. Yeah, man. When I started like uh my business, I was questioned like how could you how could you charge 10K for a website? Like it's is it like even possible? Like a friend of mine used to ask me that 10k in coffee consumption to make it. Yeah, that is churn it out. Get it done. That is what you need. You just don't sleep, right?

Yeah. Yeah. Okay. Yeah. uh at the end let us know what you think like yeah do you think like Marcel actually uh charged 50k and if they did for what Marcel if you're watching this can you just specify I don't doubt that you charge 50k but can you specify for what exactly that would be uh nice maybe sometime maybe you can show us what uh was the build so uh it would be nice to see the landing page in action and yeah and Brett if you're watching, what were you doubting in that original tech, like in the original tweet? Like what what caused you to be so dumbfounded that you couldn't possibly have charged for 50k for?

Um, are you just focusing on design? DMs are always open. DMs are always open and we don't block you guys. No, that's so that's so stupid and petty. Oh god, cringe. I don't know. I like absolute cringe. Yeah. Anyway, anyway, Twitter drama site of the week. We got to blitz through this. We've you've natted on so much on the first two things. We're not even really in the show yet. Yeah, we are not. Okay. So, for the site of the week, I don't I can't use my uh that browser. I need to use Chrome. Oh, that's not a good sign. What browser you? I was using um Arc, but I'm going to use Chrome because there is a nice thing about this website that I will show you uh that Chrome uh can show but copy link but uh I can't do it in Arc because Arc uh popup window works differently or it doesn't have a popup window at all.

Okay. So, I'll reload the site once I'm ready. Okay. So, I'll show you this. Like, oh my god, this this just so cool. All right. So, this is site of the week and this is probably the best mix of art with technology I've seen so far. This is from Shopify editions winter 26. And this is absolutely amazing. Just look at these transitions. And look at this like WebGL effects when I scroll. Do you see also like the the parallax and the art, the mix of the coffee here, the shopping bags, you know, the the Shopify thing, the skateboard, the shoes, and the Renaissance. And by the way, uh that you see what they did there, the Reni sons, the Ren AI sense edition.

Yeah, they that's nice. The AI edition. And man, the motion here is amazing. The the art is just I freaking love it. The mix that they have here, they have like 3D elements. This is like an actual like 3D element. And then these are kind of like 2 and a halfD. You see like it's it gives perspective and this is like a video somehow or like it has some motion in it. It's just so good. So they talk about a bunch of things that happened in the last year and or in the current year with Shopify and I believe like things that is going to happen. So they have this sidekick. Um then you you go through this.

By the way, these are also all animated. I'll just fly through them. Uh they have like this like really cool UI that you can hover and see and you can click and go and read, by the way. So, a lot of this is for reading. So, this is essentially like a magazine and then linked to places to to read a bunch of stuff. And then this transition breaks me every single time. It is so freaking good. It's so smooth. It's insane. And like do you see that? Like the level of the details they have the in the imagery uh with the Yeah. And and again you see this 3D book in a in a painting.

Uh it's just so good. Or this one. This is one of my favorites. I mean you see like the code that the the art on one side the code on the other side and the the motion that exists without me doing anything and that uh not to me I I forgot to mention like one of the best things about this website is the font pairing. So you see this this font that works really well with the imagery and then they have this one that is bold. It's just like the the contrast of serif and sans serif is just so so wellm made. Um and yeah, one thing I was uh thinking about like looking at these uh imagery they like I can see that um it is not too difficult to create something like this with AI like just the image like just the image with AI do does it really well.

What they've done really well is that they they pushed it enough that it doesn't look like AI generated imagery. It is just alive and has details and it has craft in it and it makes me up. It's It's so good. I look at this. Like this is amazing. It's just insane. Yeah. And and then they have this they have this like whole 3D thing and you see how the background is moving. It's just so wellmade. And then this desk in this beautiful place and yeah man it is it is out of this world. Um like the whole I I see the whole Twitter talking about this. I will make a short video and put it on all social media about this in my way of like you know with my phone.

I will make a video about this. This is just too good. And it is also very readable. The way they've done this it is extremely readable. like I don't have an issue reading anything uh even though there are so much 3D and imagery going on. Maybe here is not as readable but it is still quite readable. So they have this I mean look at this still and the globe you can you can play with the globe. Uh and then so here we have this one shop app uh the B2B. It's also really like fun how they did this like the the B2B the image. Uh you see and these are also 3D because you see them like move and the they reflect the the light and then the reason I use chrome is this there's a nice thing about it.

So you see the shipping and you see what they are doing like the cheetah and being fast with the shipping but still using the same like environment and concepts. So the the cheetah has the this box um for it's called cheetah fleet. So the details here it's just so cool the cheetah fleet. And you see this key I can click on it and look what happens. A new window pops up that is in sync with the window in the background. This is like insane. Oh I I got it. Yeah. So this is really cool. So you you take the key kind of it's like grabbing the key and now you're moving it to somewhere to open something.

So here we can click it and then something happens in this screen and in the back backgrounds. It's just insane. You see how the the art style shifts and it's kind of like a a new view. Now you can go back the whole website and see a whole layer on top of what you could already see. So you see it is it is insane. It is so it's like a detail on top of it and and then we end with the developer. This is also one of my favorite one like with the finger and these things coming together. Yeah. Um, so for people who don't know how websites like these are made, you should check out WebGL and 3JS.

Um, there are like some really good 3JS courses. Uh, the one from Brono Simon, we'll talk about that next week. Um, but yeah, you can check those. I want to share one last thing about this website. uh or like in general, Shopify has this thing that this Shopify edition that they do every year. And I want to to compare this is from this year, right? The the winter 26. I want to show the one that they did the first one that was summer 2022. Uh and what is really interesting to me like how far they have came in in terms of like graphics and like details to add because this is in concept very similar to the other one.

See, like we have this sticky navigation and a bunch of like information to go through with some imagery and some like I don't know 3D but this is what they have they had in 2022 and yeah it is you you can see the concept like the wireframe is essentially the same but man they've came a far like this is just so good what they have now. So yeah, uh check it out if you're interested in uh Shopify. You can also go to see all editions. It's in a fun magazine like uh this is where you see all of them. The other ones are also incredible like the especially the last five. Incredible. Um yeah. Anything you want to talk about this website?

There was a lot there. Yeah. No, I think I think this one's really nice. What was it? The Was it your fans I could hear spinning up whilst you're on that website, though? Sorry. Was it your fans I could hear? I don't know. I You should get your GP little monitor just to Oh, man. I can hear the Because I have the headphones. I couldn't I couldn't hear my fans. Yes. Yes. Well, I have two instance of it uh open to be fair. You're just reckless, aren't you? You're just reckless. Uh, no. I mean, you know, it's uh it's cool. What does the mobile version look like? I haven't checked it, but let's see. Let's see.

Um, so yeah, it seems like it's so cool. It's the same. I mean, again, I don't want to crash your computer, but it does look pretty pword. Yeah. Uh I don't know. I I I think it is still readable. It's just that uh the text is taking more um maybe this is not the exact representation of how it Well, screen that you'd have to look at. Yeah. Looks But I I think it is pretty good. I I want to talk about one last thing that I noticed and I don't know how they h how they did this or like also the attention to the detail. You see how I have this sidebar here, right? So if I close the sidebar and open it, you see this menu, it animates like Okay.

So every time on window size, window resize, they're animating the these elements to the new position. Is that how you would do it? Yeah. Yeah. It's pretty cool like that. It's got to Yeah, you've got to you've got to um So, you see it there. You don't need to I mean the fact that you're just making the window size. What happens if you make the right hand side smaller? Has he crashed again? Guys, let me know in the comments. Has he crashed again? Because this happened last time. Kaza entered the studio. Tisk tisk tisk. So I guess you were saying something about my fans. I was saying something about your fans. And I also tried to avoid the using the P word.

This is the second time. This is embarrassing, dude. This is the second time you've done site of the week and you have crashed your ownership. I blame Safari because my computer is running smooth. My GPU is burning a comment. I I would say it it's a because I'm sharing my screens through Safari because Chrome can't handle it. Uh can't handle our streaming service. It's this maximalist attitude. That's what it is. Two instances of this website running. I bet how many tabs you got open on the right left hand side there. Scroll down. Quite a few. Like quite a few. But I have like This is one of the many spaces that I have. And I have Okay.

So, it's just one of many spaces. Okay. Yes. And I have two Figma instances with many tabs. And I have cursor and you you know the meme everything is fine. Yeah, I know the main that poor doggy thing. Yeah, this is fine. Nothing is warning. Nothing is wrong. Okay. Okay. You do to yourself. You do. So let's um let me remove the this and also we can end it at everything is fine. Cool. Should we move on to GBT5? Yes. Yes. Yeah. Uh so close those instances without the risk of sounding you know extremist. Open eye. I have one. Project garlic scores 100% in arc AGI. But what does it all mean? I suppose I didn't go back over that one and uh change the question, but um yeah, after massive response from Gemini 3, I've lost the tab that we are on.

Um, code red being called on at the open air band camp. I really can't find this tab. Where the hell is this tab? React. I've lost it. Oh, it's because the uh the icon changes to a little red square on on on comet and I never know which uh that that's the tab shade. Anyway, sorry. So, GBT 5.2 to thinking as scoring 100%. And by the way, guys, I'm not like a benchmarks are the bane of all of our existence. It's all comes down to real world usage and what you get on with. They're a nice indicator, but we never we never fall trapped to thinking, "Oh my god, it's that amazing." But it's hard to ignore.

The one that I find hard that's very hard to ignore is the SW bench overtaking Opus on the SW bench. Now, I've been using Chachi BT5.2 today to today, guys. And my god, is it slow. My golly gosh, is it slow. Through the API or where? Uh through the API. It's a weird sound. Uh through the API. Um, and uh through warp terminal, which is my terminal I use. I set it on a pretty hard task. And because here's the thing, we talked about this last week. Uh, when I and I talked about it in my video about running local models, right? A lot of these benchmarks, a lot of these oneshot things, they throw in a single page app or whatever it is.

they just they're not very like and but once you put it into a code base where it's got to read loads of context that's when they start to to falter saying all that I gave the GPT 5.2 a task where it was just like go do this one little task don't worry about the rest of the codebase honestly I fixed about four or five bugs using Opus um in the time that it was running and running and running and then it still didn't fix it still didn't fix the bug that it was trying to to figure out. So, you know, it is what it is. But the real the real question which we've been talking about recently is like does it does it actually even matter?

Like the everyday user a f I think they're not really see people are still stuck on GPT40. They want 40 back. They don't care. They're not caring about the tone was better. The tone. Yeah. Exactly right. Um what makes a what makes a code what makes a model good is not these benchmarks and not these numbers. It's the real world usage and people still are more than happy with 40. Um so whilst open and I are barking up this tree of like AGI and the best model, people are still they're like we have a is it like a saying or something along like the line uh the lines when when the moment you start to have a test or like a benchmark it's obsolete.

Well, yeah. I mean that you know there's there's an argument to say that um you know they're optimizing specifically for these benchmarks but the thing is I see it all over YouTube you know currently is that people us you I the viewers all of us we look for that one magic pill that just tells us whether something's good or bad when we we've sort of lost the ability to even consider that actually things need a little bit of work a little bit of time a little bit of adjustment but we're we're just we're addicted to that oneshot prompt. We're addicted to those sort of bombastic headlines to say this is, you know, the best or this is the worst or whatever.

We we're losing that ability for nuance. Well, we've lost it completely. Do you know what I mean? Um, but yeah, there is an argument to say that the they are they are optimizing for these benchmarks because I still showed you the benchmarks even though I have this opinion that they don't even matter. Have you done any of the benchmarks yourself? The RKGI ones? No. No. No. Like there are games you can you can play them and there there are like no instructions so you have to do something and I'm not going to spoil it. I've done like a few of them. Um and they can be tricky sometimes but what's your score? 20% less than GPT.

Now I don't know. I don't know. I I never got a score. You just get or I maybe I haven't done enough of it but like you you go to the game and there are like levels. So level one is simple and then you figure it out and then you you advance the levels. Um but you get a feeling for it like may maybe that is the human part like okay so I need to rotate the block or I need to to to interact with that color but not with that color. So it's that type of things and I I guess for AI it takes many runs that that's that's what I what my understanding is.

It takes many runs to get all that necessary information. Okay, this is how I interact with this game. Um when while for us humans it's probably just order of magnitude easier because we are very efficient. Right. Right. We're like our brains are like incredibly efficient. And the amount of data that we take uh visually and through sound and everything else is just like insanely uh huge and the computation that we do at the same time. So what what is uh interesting also about these um models getting better at these um AGI at at these like uh what what are they called again? the uh the RKGI tests uh is the cost because it used to take like thousand and thousands of dollars to just do one test and now it's just uh much much less I believe uh it was like Theo mentioned like $11 or something like this something that was in thousands.

Well, funny you say that because I wanted to just raise the price real quick, which is um 21 in uh Hang on. No, that's 5.2. Should I Should we share the screen? Uh yeah. Yeah, bring it on. Bring it on. Um if we go to 5.2, uh it's 175 in, $14 out. So, it's it's gone up in price. Um but saying all that, you mentioned Theo. Um he like he he released it and I was in the comments section with people and people are like tired. People are like h like whatever like you know and and on the one hand I get it. Um so I've got a couple of videos on LM Studio if you look on my channel which is there that little thing over the side there 0x5 and5.

Um or just search Samuel Gregory. Um, and people are tired, right? And they're they're exhausted about all these these new models that come out. And it's I said it surprises me because we're at the cutting edge of technology. We're supposed to be excited about new technology. And look at how stagnant Apple was. And I know, you know, we're comparing apples and oranges potentially, but when Apple just weren't releasing any meaningful updates to any of their computers and then people complained about that because there were no updates. Like I I I for sure think we should celebrate any and all improvements to technology. However, in the name of crossing over a thousand subscribers on our on our YouTube, right?

What do you want to hear about these new models? Like do you you know balancing that we should be celebrating progress and we do want things to get better. Um rather than just kind of like dismissing it because it's just you're not the new model. I would like to make these sections more interesting and more kind of like, you know, maybe it's more anecdotes. Maybe you just don't want to hear about them at all because as whatever. One guy was like, "Wake me when when GBT7 is out sort of thing." And and on the one hand, I get it, but on the other hand, it's like any progress on any any progress is should be celebrated.

Like I don't think we should just silently let things, you know, um silently let things uh progress and just not not celebrate that. Basically, I've just said the same thing like seven or eight times, but I don't know. Let us know what you want to hear about these new models. Um but whatever. Uh sorry, Dominico, I I brushed over this sort of stuff. Yes. So, LM Studio, I've released two videos on my account. So I downloaded um um Quen 3 coder and I downloaded uh Devstrol and I set them up to run with cursor and do my local models and and whatever because um but there's a video that I did that I think I did a pretty good job of like breaking down what you need to look for when it comes to these models because again when they've got small context windows, small context size, you can't plug that into your code base because it just runs out of space too quickly.

Unless there's something that I'm missing and I've asked the open question in the video. Um but yeah, otherwise LM Studio you can just do basic chat with it and use open source models and stuff like that. You can also use Lama. Um but yeah, let us know what you want to do with it. But I've done some pretty extensive testing with it. I even got it running on clawed code as well. So, I use I use Claude Code router and LM Studio to use LM Studio on um local models. All I'll say is I've got an M1 Max and it took ages to run this stuff. If you've got anything kind of newer, I'd love to know what sort of results you're getting, but it took ages.

And I think generally we are because we're competing with H100s and the server that are running by these very very big um corporations, you know, so it's hard to beat those sort that sort of performance. So yeah, anyway, React uh React uh GBT 5.2 I'm here for it, but not really going to affect my dayto-day. Uh especially in coding. I don't know. I don't know what what what they're going for at this point because the normies are happy with 40. So yeah, let us know if what you're doing with chat GBT, whether you use it for anything kind of cool, but they're just they're just a meh company for me. Do you have anything to add to that?

No, my favorite model so far has been 01 is it? Is that really or just 01? Yeah, like the best vibe coding I've done uh was with that like the best results that I've gotten. Maybe uh also on a note on what you ask the audience, maybe we could have like our own um kind of like test since we are a bit more like doing things like you know we're talking about the web space stuff. So, anything with web design and development, maybe we could figure out a test that we could do like maybe you know that um landing page that I have for flowing.2 with that ribbon and it's like multiple stages and it's quite complex.

maybe trying to see again and again uh how close a mo model can get uh to creating something like this with one prompt I don't know two props it just wouldn't come from an authentic place for me though just because I don't believe you can judge a model based on these oneshot benchmarkes things right um so how can we judge a model yeah well yeah exactly you have to put in the time you have to put you have to use them. I watched the uh I think it's Theo's um channel manager also has a YouTube channel and he he hit the nail on the head at the end of his video on his setup because I want to release a video on my setup as well actually.

He says the real test is just sitting down and playing with them. Like we can't be looking for these these magic pills that just tell us everything we need to know. Like we we're getting lazy to be honest. I think we're just getting too lazy. These models are very non-deterministic. So we have to have a way to compare them and do it with repetition. Like we can't do one prompt, right? just it's just fundamentally impossible to like Yeah, we're deal and he and and Ben this guy he says this in his video he says we can't because they're so undeterminist indeterministic undeterministic because they're so indeterministic like it it further further exemplifies your your everyday usage the way that you like to use code the right the the way that you like to use things you know.

Yeah. And the point is with further prompting would be based on the the output. And because the output is nondeterministic, the the prompts that you give it um brings you further and further away from the original thing and makes it infinitely more difficult to compare them the more you prompt. So each prompt is exponentially adding to the non-deterministic nature of it. So it's and so it leads us to nothing uh other than these like um benchmarks that are useless. Sure. Yeah. Exactly. So which brings me to my point is how can we bring this news to you in a valuable way? What do you want to hear? Because yeah benchmarks are useless. Um I can give you anecdotes uh on my personal usage which again may or may differ may differ from your perception.

You might like its overall attitude. Um on the other side it's hard to switch from one model provider to another just to see it performance better especially if you're working. Yeah and I agree. So I was looking at um codeex because of GPT5 and I want to give GPT a go you know but I'm I'm you know just thinking about all the agents that I have set up in claw code all of the rules it's not that hard but I I totally understand what you mean it's like you know and that's what these companies want that's why Apple put up these wall gardens that's why web flow put put these wall gardens and that's why Android make it just a little bit more difficult to kind of move around and whatever.

Just a little bit of friction or the fear of of moving around is enough to stop people doing it. Um, and and I go back to something I said last week, like this time last year, maybe a little bit before, we're all handwriting code and now we're like splitting hairs on the differences between one model and the next. Like we are spoiled to say the least, right? So at the end of the day, if you're happy, if you have no need to search for another tool, stick with what you stick with what you know, use it, and then come back in a year's time where where there might be some meaningful updates. But um if you just don't have the energy, time, or whatever to be looking at new models, why bother clicking on a video that reviews the new Chat GBT 5.2?

Like, and we were talking about this before the before the show as well, how how um eccentric YouTube videos have got. They make you believe that 5.2 will literally change your mind, impregnate you, even if you're a guy, and literally set the world on fire because it's that good. And it's just not the case. And these YouTubers are really, really selling something that's really just not worth getting hyped up about. But I don't know. Um, I don't know what you Oh, um, chat GBT to extract data from a doc, turn it into piece of CSV. Yeah, and that's straightforward. I don't think you need the latest model to even kind of do all that kind of stuff, you know.

Um, it's pretty pretty standard behavior. Sorry guys, I'm reading this this comment here. Uh, what what Dominico is doing with his chat usage, splitting up docs. Um, I mean, to be fair, they'll take two seconds to whack into Claude, right? That's not like it's not like you've got a whole ecosystem around it. But again, if if it ain't broke, don't fix it. If it's getting the job done, stick with what you're using. I guess uh what you can do while your AI is working for you is to test other AIs. Yeah. Yeah. Yeah. That's another thing, man. Like, you know, we need to learn to chill out a little bit. Like AI is doing a lot for us, yet we try and fill our time with like more stuff.

It's like just just chill, you know. Oh, anyway, whatever. Too many um too many whatever. We'll talk we'll talk about React. Uh the worst thing Vibe coders could ever have not had the foggiest idea about has happened. There's a new React server component exploit that's been given the CVSS score of 10 and is on fire. So, that is my way of saying that this this is actually pretty a pretty big deal. So 10 is it's 10 out of 10, right? So it's the worst critical. Yeah. Yeah. You can run you can run um the server is the safest place to be. It's where you store your environment variables. It's where you run code that means that people are authenticated that they um it's it's really like the last line of defense which should be bulletproof.

yet with an exploit that React um have exploited revealed on um their server components. You can run server actions with zero permissions like you can do anything. You can expose environment variables, you can contact databases, you can do all this. And the nerdiness of it is basically it's in JavaScript everything is an object. Everything is an object. Even functions are objects. Objects are objects obviously, but everything boils down to an object. And objects have a hidden. In fact, if we were to uh inspect um I don't know how we do this. Uh console log. I want to try and make I want to try and make an object. So maybe if we go const try and bring that up here.

Um, const object equals new object. Oops. Object. And then we console log dot OBJ. So, you see this? Can you see that? It's like got this. Uh, we still see the Oh, sorry, guys. Sorry, guys. Uh, how would I do this now? Um, can I maybe I will share the window? Maybe I'll share the window. Yeah. Stop. Uh, share screen window. Boom. Let's do that. So, what I did there was I just created a new object, right? And then console log the object. And automatically it's got this prototype. And then inside of here, you've got all of this secret stuff. And this some of this stuff or some of this um and it basically exploits the the fact of this prototype that you can then run other you can access other bits of code on the server through this.

So there was uh the get the get method they were able to call the secret get method on on um objects that can that have data stored in them. So, you know, this is all incredibly like nerdy, but just to know that this is kind of how they've exploited it. And uh if you're running anything um anything on React, anything next.js, whatever, update update to the latest latest version that you can. Uh really simple. You can ask claw code to do it. And it even worked out that payload haven't updated to the latest version of React. So it's the patch version that you want to want to update to. And then on top of that within the last day um remove they found two more exploits um that you need to update to.

So if you updated based on this bloody this this exploit that's just happened, they found two more. Um, one of them was a denial of service uh exploit. The other one was source code exposure um in this. So it's like yeah it's is on fire. It's pretty bad. It's pretty bad. The the denial of service one is an also is is I don't think it's a 10, but it is pretty it's critical. Denial of service. Here we go. Uh okay. So no 3 7.5. There you go. So, oh, sorry, I'm not sharing my screen. Still pretty bad. There you go. 7 7.5. So, man, and this has been, by the way, this has been in React since day dot.

So, they've got to they've got to provide this patch on all of the versions of React. Actually, no, has it been day dot or maybe day dot of server components, which is like a year or two old or whatever. So, there's a lot of um there's going to be a lot of websites out there that use this version. So, yeah, pretty pretty big deal. And uh if you're running any websites on React, I would I would just update to the latest patch version. So take your version and just add a add a number to the third digit onto it and should fix it. But yeah, pretty pretty crazy. And it goes into what you were saying as well about like software just becoming absolute garbage at the moment.

Yeah. Like yeah, everything like my browser crashed, mate. I I don't know about I don't know you can say anything about your uh your crashing. I might have a thousand tabs but my computer should be able to handle that. We are talking about HTML. I don't know. I feel like we we see with this uh level of you know carelessness with AI like software will just get more and more trash. M we will have less optimization in most things and yeah um I you know I speculated it's it's in part due to the pressures that developers being put under deliver fast. That is true. Yes. you know, quality assurance cuts on um QA testing and things like that.

Like we're just it's gone from, you know, agile was this idea of delivering value quickly in order to learn and bring that back into the into the the tool or the product to learn what actually people want. But you still need a robust testing process to go along with that. um and cuts are being made and all the rest of that maybe this information or this this this commitment or dedication to quality is just getting lost. But I just I just don't know. I just do not know. Um I'm just going to Oh, wow. I've got a video feed open. So maybe that's why I was lagging apparently. Um should we talk about uh comments?

Yeah, the comment section. Do we have uh we have comments on the I believe like on the current web flow situation. Yeah. So basically we uh well this is the comment section where we take a look about what you guys have to say. Let us know what you think about anything you've heard or whatever in the comments and we'll we'll we'll try and take a look at it. But um the latest video that the the last week we spoke about Web Flow needing more creators. They seem to be disappearing, you know. Um, and we're getting a lot of commentary kind of agreeing really even though cuz we we spoke about some new cool tools that being released by some content creators overall the content is just far less and I don't know whether I'm in in a certain um uh algorithmic bubble but I'm just I'm just not seeing much in the way of web flow.

I mean a lot of people here are us saying about the product position being changed. uh they they're more guided towards enterprise clients which I do think is is a big thing. um like that's I think that's just but what I don't understand there um like trying to see it from web flow's perspective why would you sacrifice the community for going towards enterprise you can go towards the enterprise and still have the community that's my issue like even um if you want to be an enterprise tool you want people to to choose your tool and build with your tool and that is your community. So you nurture them, you help them uh you want them to succeed because they use your tool to in this case towards enterprise and yeah that is uh a big part.

Another one like the se second comment about framer sponsoring everything which is also true. Framer is sponsoring everyone and it looks like everybody is trying to make content for Framer. They also uh they they gave away 100% of the revenue from templates to their uh creators. So as a creator if you would make a framer if you still if you make a framer template you get 100% of that money. that also made it so that many people started making templates and sell it and just keep the money and talk about it. Uh I think that is also true. That is a big one. Um and yeah. Yeah. I I agree. I agree. Um people are jumping ship.

I think that was one of the I just wanted to align what we're talking about here as well just because you know what uh Whiskey 88838 rolls off the tongue uh is talking about here is is is what Web Flow are doing you know and this this does play a part in the conversation I'm happy to chat about it but like um I think my my commentary is more about the community creating content for the rest of the community irrespective of Web Flow's goals and where they're directing their attention to like this this no doubt will have a knock-on effect because maybe maybe there's no features to talk about because they're only releasing those features to um to to their enterprise clients or whatever.

They did however just release today they've gaed um the the web flow cloud storage stuff which has been in beta for for a while and I you know I'm thinking about making content on it and I've got a course on the content but to be really frank it's not performing great and it's very hard to for me to justify time to create create tutorials that just aren't people aren't seeing or slash people aren't watching people aren't enjoying And I wonder why people aren't enjoying them as much because the con the content is being created, but I'm not getting the watch hours that I used to. People do like the yearly stuff. I'll probably will do a yearly video of, you know, the breakdown of Web Flow Conf.

But that just seems like I'm just cashing in on the stuff that's doing well, but it's it's guys, it's a two-way street. You know, we I've got to be energized to do stuff that I'm passionate about. Um, and also that you got it's a marriage, isn't it? a ven diagram of people what people like to watch and what you enjoy doing. So, um but yeah, um I did see uh Timothy Ricks might be still kind of leading the charge and that's what makes me think I'm I'm in a bit of an algorithmic bubble a little bit because I'm not I'm not getting them on my feed. Um but uh what else have we got here as well?

like yeah well just on that alignment thing I just wanted to bring it back to like why are community members not producing content and um how that plays into this this enterprise client direction that web flow are going along to um content creators hype machines I mean I don't necessarily agree with that they are hype machines I think that again it's that ven diagram right we've got to we've got to balance what we enjoy talking about and also what stuff you want to watch. So, I don't think they're constantly chasing the hype. Um, and this is this is the comment I wanted to get to. Uh, no one and it's it basically echoes exactly what you say, Kabaza.

No one really knows where Web Flow is heading anymore. Features announced then dropped, memberships gone, e-commerce left behind, random AI bits without a clear plan, then price hikes push forward and a push with enterprise. Um, I think downtime is probably to do with Cloudflare, not Web Flow. But yeah, people don't necessarily know that. They just see my Web Flow website as down or whatever. No, actually, no, they had two, didn't they? I had two a couple of months ago where it just kind of and it was I think it was to do with Cloudflare, but either way, it was they they took a lot of responsibility for that for whatever reason. But, um, the competition is there and people aren't, you know, people aren't watching the content that's being put out because they are distracted by other things.

So I don't know I don't know what the conclusion is on this. I disagree with the part that uh there aren't enough demand or like same demand for web flow developers. I I believe that is still growing. More and more companies are using web flow and there is more. Who's saying that there's no demand for web flow developers? It was in the same comment was it um still platform still flow where? Oh yeah, because the demand for web flow developers isn't what it used to be. Um, yeah. No, I disagree with that too. I think a lot of business the the platform is absolutely growing. What we are talking about is not like the the platform not doing well.

It's that that aspect of caring for the community and like also Web Flow used to kind of lead with Webflow University. said the community was like very excited about creating content but that has changed. So, but um it is what it is in a way. It is what it is. Uh I I guess they are trying to get back uh on track, but we'll see. We'll see. Um and just a quick shout out to someone we've had on the show. I don't know whether you guys have seen this. This is probably before we had even had this separate channel, right? when it was on my channel, but um Rhymar um who's a channel friend of ours just been hired by Replet.

He's been he was he was letting me know what was going on behind the scenes. So, I did have a I did know this was happening, but he's finally he's finally been hired and this has honestly been a dream of his. Like, again, offline him and I nerded out over Replet. We spoke about Replet on the show. So if you don't know what Replet is, it's one of these I think more better vibe code platforms that does a really really good job of building really robust and dare I say enterprise level um uh apps, you know, coded apps. Funny Connor Connor Finles is here and that was one of the comments that someone misses Connor because he's not producing stuff.

He's he's doing stuff on I saw a video him doing stuff with Memberstack. I think I think he's doubling down on that or whatever. He's kind of like I don't know. I think I haven't seen much from him for a while. But again, he pivoted away from Web Flow. I pivoted away from Web Flow. Um but yeah, um there was this content creator called Kabaza a little while ago that was uh doing Web Flow content, but I haven't seen him for a while either. Yeah. He's he's evaporating on his community on his product and trying to launch that. Yeah. Yeah. Yeah. Got to go where the money is and exactly and even with the product that I'm trying to launch uh we started with just trying to do web flow stuff and now we are trying to do not just web flow but you know help anyone.

So that is also you know showing that I can't f I can't count on just focusing on web flow. But very happy for Rymer and uh very welld deserved. Yeah. Yeah. Really really cool. He's one of the reasons why I got into web flow. Ah yeah that's nice. Yeah it's nice. And we had that old debate about word WordPress versus web flow on Facebook. It was there's that was a money earner video for a while that was always being like recommended on any kind of research towards like web versus WordPress. Yeah. Um still the case like loads of comments on our videos are talking about um WordPress people going back to WordPress you know I I don't know like to me WordPress doesn't exist anymore.

We miss time and place. Um yeah, we miss Magguire. Absolutely, man. He was h he was just so good. He Yeah, what should I do now? He did he he started a a startup around AI and teaching and I don't know he he wanted to push content weekly and update us but then he stopped. So I don't know. Oh yeah, I remember that. Yeah. Yeah. Yeah, he was really good. So a learning for like any team, any company. Um you will have someone like Maguire like a north start like amazing but you have to build systems around them. So because they probably will leave eventually that you don't lose everything that you can still use the systems to to somehow continue.

Uh we have another comment. Replet is amazing and and that was a great hire for them. Yeah. Yeah. He's big on community. I think he lives and breathes community like um Romar like even offline he would just talk about he fundamentally believes that it's the secret source to like most aspects of of a business which arguably he's right you know um the community word of the mouth right yeah exactly exactly yeah that's exactly what it is so um I'll take a moment to say by the way if you are watching on YouTube give us a like smash the like button subscribe because we are like closing on 10k by before the end of the year.

Yeah. Yeah. And uh we really appreciate all your support in the in the last probably less than six months. We've we've been doing well and we have some really really cool um I mean we we still got visual updates in cursor. I don't know whether you guys saw that visual updates in cursor and some new fe Figma features as well. So, we'll blast through these next points and then we'll get to the juicy stuff. But, um to uh what's the new channel again for us on X? Is this Can you Can you screenshot your screen with the QR code uh in the top left there or top right? I don't know what side it's on for you.

Command AI show is the is what you cmd. Let me see if I can do this. you what the screenshot thing. Yeah. Yeah. You might have to save the photo, but you might be able to do it from the from the screenshot screen, which is so Oh my god, I'm hating liquid glass. Like Yeah. Uh we we didn't cover this, by the way, since you mentioned it. Let's talk about it briefly. The guy at Apple who was responsible for Apple for this style the glass of he he left for Meta for Facebook. Um and they didn't expect this because Apple does not show someone in their you know videos knowing that they will leave. So it caught them by surprise.

But what is really interesting like there were some there were some leaked uh comments on him leaving. People at Apple were really really happy. There were people celebrating because they didn't really like him. The guy I thought he was hardware though. I thought it was hardware design not software. No, he had a f he had a background in fashion and he did the the liquid glass stuff that is also like physical world bringing it here and it does like yes you you need to get inspired by the physical world for your UI but not literally not this way not you know sacrificing so much so end of story people are really happy at Apple that he left and they are looking forward like I am looking forward for the next generation of um iOS and the new designs moving away from liquid glass hopefully and just I don't know it's like it's it's it's kind of less about liquid glass saying that I turned down because you got an option now to like chill out on the liquid glass stuff it's more about like the actual implementation of the of of the operating system like um like the mail app just does not work properly like I don't know how I would find email if I didn't have a a keyboard attached to it.

Cuz whenever you just go on it, the the the search box is like disabled. So I have to press command F to get it to work. And then that's my way of finding stuff. Little things like that. When I open up a new mail, it pops up over to to the left hand side of the screen and then just jumps to the center. And it's just all these little things that have nothing to do with liquid glass. Although the stress and whatever of having to do everything app for example as well. Yeah, it's just it's crazy. Yeah, it was good. The photos app was good and they ruined it. Now it's again tolerable but it was good and then they changed it and made it worse and then yeah.

Um anyway, that guy is gone. Gone to meta. Um nice one. Thanks for subscribing and uh thank you appreciate it. Speaking of uh tolerance and tolerating anthropic have donated the MCP pro well the model context protocol to Linux. Um it's I don't know how what to make of this to be honest whether this is like a lastditch effort to be like just just can you sort this mess out for us but um it's definitely something they've made some notable improvements on the last definitely the last few weeks with the tool search tool things like that but there's definitely a very very divisive the the ambitions of MCP if you don't know what they are it's a natural language wrapper for APIs so um if you want to update a database case, you don't have to like send it a JSON bit of code and all this.

You just say update the name and address to this and it will figure out everything it needs to to do to interact with that API. Um, but yeah, celebrating all this success all the way down to the Linux Foundation. It's a new agentic AI foundation that has actually been started by anthropic block open with with support from Google, Microsoft, and AWS and Bloomberg. bit of a random one there. Um, and yeah, they've sort of donated it uh to it's basically essence open source now and they'll look after it and and the AI Agentic Found the Agentic AI the Agentic Artificial Intelligence Foundation uh will join by Goose and the Agents MD. I love agents MD.

I wish claw code would adopt agents MD. It's just a a a standardized format to give your AI agent certain rules. But yeah, um and so we'll see whether they make some whether the community can kind of fix it or or help it or whether some people outside of Claude's little weird bubble that they have um can sort of breathe some life into it. But, you know, I think a lot of devs are really reconsidering MCPS just because they're just a pain in the ass to manage. Yeah. And how does this um work like the foundation accepts this obviously like you can't just say okay I have this trash I'm donating it to you right so so there there has been talks and the the foundation has to accept it and like do they have to give like any assurance on if they will uh continue developing if they will continue supporting and things like this what if there are like I don't understand how that part works but also oh go on like on the same note essentially what does it mean um what does this mean for the MCP protocol if um the Linux is taking it in the foundation does it mean that they believe in it right that that to me signals something more important than the whole thing.

Uh potentially if they believe in it and they think this is the future that that might be interesting to talk about. Yeah, potentially. I mean at the end of the day this means that the community will be able to drive the progress of the protocol, right? And it gives it a platform in which to be able to support that. Maybe it's funding it. Um you know, money being brought in by the the uh because it's probably a charity or something. Um it's yet to be seen what comes out of this and it's yet to be seen exactly why they've why they have donated and why they've accepted it while the AI foundation have accepted it.

It's an interesting move. I think it's a positive move. Um because again, you know, Claude Anthropic have just been throwing everything at it felt like there were three weeks last week where every week we had some sort of update, some sort of blog post, some sort of admittance to like the the difficulty of MCP. So, it's in a bit of a weird spot. Um yeah, but it basically piggybacks tool use, right? So by um by definitions AIs or when you call an AI, you can give it a bunch of tools that it can use it. You know, the the simplest one is for it to be able to tell the time, you know, because an AI does not know what the date and time is.

So you give it the tool to be able to do that gives it extra capabilities. MCPS are an extension off of that idea for them to be able to call live data. You know, it's again it's that API aspect to it. So the idea is certainly good and and more and more AIs are getting trained on tool use. It's just whether MCP is the is the best way to do it. Um I don't know. I've got a a MCPs running all of the time. Um but there are times where again I switch them on and I accidentally leave them on and then I restart a new session and then they're all turned on again. It's like sake.

I did it today. I was like jam. I'm like trying to work and it kept on like compacting my conversation. I was like why is this guy compacting the conversation all the time slashmcp all of my MCPs are turned on. It's like Jesus. So I turned bunch of them off. So yeah definitely um and I want to try and use skills. I want to try and leverage skills but uh anyway MCPs let's uh let's move on. Um quick one on Apple. They've released I'm not sure whe this was a video from a uh another com another YouTube channel called better stack. So I don't actually know what the timeline is on this but it was interesting to nonetheless that Apple have released a opensource image generation model code name Zulu is asking what MCPs are.

Uh well actually no I just explained to you uh there are tools that it's a it's an a it's a it's a wrapper around APIs that you can give to an AI yeah wrapper around APIs that you can give to AIS to be able to access live data that's that's in essence what it is um but yeah Starflow um and it's an image generation thing which by the way I tried setting up before this show but I couldn't I don't even think it works on Mac. So, it's developed by Apple, but it's actually just something that can probably just Linux or probably something on PCs or something, but it doesn't work. Anyway, that's not really the point.

The point is is that they've released this open source image generation model which uses something called the novel transformer auto reggressive flow architecture. Okay. Yes, you heard that right. But it's a new way to an efficient way to generate images. So I think the bigger question is here is like one is open source. Two is Apple, three is AI. Never did I think I'd hear those three things in the same sentence to be honest. Uh but there here we are. We're with Starflow and you can go and download it and use it now if you've got a PC. You don't you don't have a PC? I don't have a PC to be fair. It could be something I'm setting up.

I I basically just followed these instructions here. Um, and uh, and it was just, you know, he's even got a setup script. So, you just run the setup script and it should do everything for you. But it's not. And I don't I don't know why. Do you think I can do it? Chow. Chow. Dominica contributions. Sorry. Do you think it's simple enough that I could do it? Yeah. Look, this is it. And then and then you run and then you run that with a command, a cat playing a piano, and then it's done. Or then you can break it into like, you know, you can play around with the images are tiny. I think they're like 256 x 256.

So they're tiny images that it can generate. But um I saw someone in the comments of the video that I watched and at the end of the day the the the quality isn't great, you know. So which is why they've they've released one. Oh, Claude has a contributor. some vibe coding on it. Uh, you know, at the end of the day, the results speak to themselves. It's not a great it's not it's not a great um images, but the progress is there. We're here to celebrate progress. Um, but I wouldn't be rushing to generate your images with this yet because they're just not competing with um with with the likes of other image generation models, you know.

Nano Banana Pro. Nano Banana Pro. Did you see my video that I released uh this week? Um Claude Code on your phone? No. No, I haven't. So So I think it's a really good video. It was sponsored. It's sponsored video, but I never I never work with sponsors who I don't you know I think is a really cool idea. It's basically an app that you run on your phone that uses claw code under the hood and you can just generate apps with it and it's really good. Really good. Um, and uh, yeah, they they sort of they they sponsored a video so I would use their app. And I asked my brother like cuz here's the thing.

This is this is this is peak behind the curtains here. I was, you know, as a developer, I'm still worried about AI. I'm still worried about the future of AI as much as I'm all in on it. And I was thinking, what industries are not affected by AI? And one of those is like, what do they call it? Like white collar or blue collar? Basically, people like plumbers. and my brother who's a painting decorator, those sorts of jobs. So, well, how can I empower those types of jobs? And I was thinking, why don't I just ask my brother if he could, if he could have any app in the world that would make his job easier, what would he do?

He's a painting, decorator. He was like, well, I'd love to just to be able to just take a photo of something and then just make the change just so I can show my client their idea. You know, often he'll have a husband and a wife. One wants to paint the wall this color. Another wants to paint the wall that color. Take a photo. Generate the image. Done. So, I did that. I filmed my brother. I filmed a phone call with my brother and I did the I built the app and it works really well. I'm trying to get it on the app store now. Um I'll let you know when that's done because I'll probably just do a oneoff oneoff £3 pound50 charge for it, you know.

Um yeah, that is cool. Yeah. And it uses Nano Banana and um it works really really well. And I was able I'm just able just to sit on the toilet with my phone and add features to this app and it's just so fun. It's so fun. Why am I talking about this? I don't know. Uh I don't know either. But that is uh that is a fun one because I thought about same thing for my brother. He used to work in a garden uh automation company. So doing like stuff for the for gardens and some part of that work was to do uh planning. So I thought like why don't you like take a take the drone take a picture of like the the whole um garden and then like add trees or like bushes and stuff like this with AI.

So it's essentially the same idea, right? So just adding things to the images. And I did try it at that time with chat GPT and the results were eh not really that good. But I can imagine it being like really well with working well with Nano Banana Pro and you could add like features like drag and dropping some elements. Yeah. So it looks easier for the user. But behind the scenes it's just a prompt saying like add this to this place. Speaking of which, it look it to me it sounds like the visual editor that cursor is building. Yeah, we'll get into that. But yeah, I guess we were just talking about image generation and I used Nanab Banana 3.

Nanabanana 2.2 was the default one that it used. Um, and it it it did an all right job. It was all right, but then I was like, let's because you can it's you can SSH into the into the app itself so I can then work on it on my computer and edit it. So, we added Neno Banana 3 and the the quality was was great. Yeah, man. Uh I think they're pretty cool. I think the again these are these are I call them normies. These guy these guys never and this is why the app or the video was so close to my heart. It's like it gives people like my brother the ability to dream and make notable improvements in their lives.

Um but yeah, if you go on to my channel 0x5M5, man, I plug my channel a lot. Um, then you can watch the latest video. It's only 10 minutes, so it's it's all right. Watch it at 2x speed. But yeah, I just build this out for my brother and it's even better now. Um, so I'm just trying to launch it and and whatever. And everyone keeps telling me just to launch it and it's it's cool. I don't think I'm going to be a millionaire, don't get me wrong, but um, it was still pretty cool. Who knows? You might Who knows? Let's move on to Figma. So Figma has just become the Photoshop killer. is kill.

It's literally kill Photoshop. There are some notable updates made to our favorite web design tool that branch further into Photoshop territory. So, they've added object erase. Wait, let's have a little look at this video here. Oh, no. I'm not sharing my screen. Sorry, guys. Object erase. Um, it's probably the end of the video, isn't it? Um, extensions, isolate object, and expand image. And you know, these aren't break, you know, these aren't crazy additions. Here you go. You can extend the image and stuff. I've been using Affinity. I've been loving Affinity's extend feature, and it does a really, really good job. It just had me interested here just because we use Figma mostly for the web and app design which is vector-based not really pixelbased but I think this will they're sort of adding features here that yes we will do stuff like this for our apps for sure but this becomes slowly but surely becoming add effects as well um and all that kind of stuff it's slowly becoming I mean we do our thumbnails in here, which is traditionally a pixelbased kind of editing tool.

This probably will kill Photoshop for a lot of people just as they're adding more and more of these photo elements. This is genuinely this is one reason why I kind of wanted um affinity with the AI feature because of the object fill. If if I can somehow work uh you know justify a way to have the rest of uh Figma's features in you know in my workflow then I might move on to move on to Figma. But these are really interesting photocentric features they're adding. Yeah, we we are creating our thumbnails in like Figma and I use Figma for so many so so many things and it is becoming like more like a a tool that does it all and yeah like Photoshop it takes ages to open and then you have files you know individual files if you think about it that way like Figma has made it so much easier because you can sometimes you just do something in whatever file.

It doesn't matter. You need to do something quickly. You just open it, do it where in wherever. Uh it's something that you can't do with Figma and with uh Photoshop. And we take sometimes these things for granted, but Figma fundamentally changed the the whole process of doing things and also like you do it and then your your team can uh pick it up and continue it. Like imagine I upload something and I just tell my team to continue where I uh left and yeah with these images uh image edit editing tools we'll be just doing more of that inside of Figma. Yeah, I've used the I've used uh AI generated stuff for some of our photos.

I think there was one I remember where your hand was like cut off. you were making like a an expression and I I AI generated the rest of your hands and put it into a thumbnail. So these sorts of little quality of life things do come in handy. So even I'm not real. You're not real in one of our one of our um thumbnails anyway. But you know um available right now into in the Figma design and Figma draw uh to all use our full seat professional organization enterprise plans. So yeah pretty pretty interesting. Um yeah, let's see where it goes. But the real reason why you're here is well cursor launch a revolutionary new visual editor.

It's probably the end of web flow to be honest. Um did you want to get an example up? Did you want to play or shall I do mine? I do have a simple example. You can or we do my one. We can do both. Yeah. Well, let me Yeah, go on. Let me um present my cursor whilst you get yours up as well. Okay. So, I have mine open here. So, right here uh you see I'm building like some animation tool. This is a tool I'm using to let me actually disable the cursor the new update. So, this is what I'm building, right? I'm building like an animation tool to to do this bun this type of stuff.

and then add it to your project. Now, um if I want in the past, uh if I wanted to change the the design, the HTML and the CSS essentially, I would have to, you know, come here and just change it manually or ask AI to do it. Now, I can click this button and then I can click whatever element here. I have this div. When I once I click it, this like design looking tool uh pops up and I can literally just play with the flexbox. This is like super cool. I can do a grid. So yeah, this is really nice. And I don't know why the position is on the very top for whatever reason.

So that but whatever. Like this is like really nice. And look, this is this looks very much like what Web Flow has. Mhm. The padding, the margin uh are here. The appearance, text, color, like pretty much everything is here. You can add tokens. And what I really like about this whole setup is this design versus CSS by the way that you can also see the CSS stuff and just change them here because sometimes this UI becomes too cumbersome. Let's say I want to add a a CSS caulk like in web flow you do it in the UI and it's like you need multiple clicks to do a caulk um and then you get out of it and you know going back and forth and changing the numbers becomes like a hassle here you you can do it like in this this new layout.

So um that is also nice and obviously you have your uh elements here the they call it components. Now what is really interesting about this something that I didn't know and you pointed out I thought that when when you do these changes it's done but it's not exactly. So the way this works, this is essentially Chrome uh developer tools visualized because like if you go to your Google Chrome and check check your code, you you essentially have all of these but in different view now you have this in this visual way. But that is just changing what we are seeing here and not the actual code. It's not changing the actual code. Once we are done with playing around with it, you you click apply and then it adds those changes as a prompt as you see display from block to flex row to column blah blah blah for that element and then it goes on and does it.

So it's still prompt based which is really funny. It it is something about this to me is extremely funny that you are doing like flex alignment visually and then applying it with a pro like we are burning tokens. We are burning down trees. I know this is a joke. We we but we're burning through tokens for something that we have already done visually. But I I I didn't know why. But you you you can explain it better. You mentioned that you said like if you're using Tailwind, you don't want to change stuff in line and just randomly add CSS. It has to know to go to the source. Like if I'm changing this, I'm not just changing this class or you know this element.

This is maybe a design token coming from somewhere else. So that's why it goes to the AI and the AI knows ah okay that that's the context that's how it should be done. There you go. You explain it beautifully. You explain it beautifully. Yeah cuz like you know again it might be that um you're using another language you're use you know you're you're using a certain class naming system or something like that has it has that as you say context. So it makes sense. Um you don't want inline styles everywhere. Uh but it would be nice yeah if they can figure out a way that you're well again you don't want to run AI with every single change that you make because you might be experimenting you might be exploring right you might be like do I want it left I want it centered do I want it right all this kind of stuff you don't want to run an AI command over that but I guess the only downside here is that it is using AI which means it's open to hallucinations so like maybe they've got a system maybe they've trained their model or whatever um I'm guessing it just uses cursor or can You just select can you you can select the model that it runs on.

Uh you can select the model. I I was using composer sorry not Wait wait wait wait what? This is kind of funny because I was using max mode GPG 5.1 codeex max to change the alignment of a place. Those poor poor trees. Yeah, it reasoned for about see that that should not if you switch that to composer now. If you just undo that or or make some more changes, go to composer. Composer will just do it like that. GBT chat GBT. Sorry, it's so so long. But yeah, it's pretty it's pretty cool. Um, you know, it raises the question, who who is this for? Is this going to I mean, they mention on the Twitter thread, this is going to bring designers and developers closer together.

I actually think it will tear them apart because a developer who really just, you know, many of them don't know or care about the design. They're they're kind of encouraged now to make a little uh tweak here or there. I hope that on the enterprise plan you can turn this mode off or or switch this mode off. But um yeah, yeah, it's it's an interesting feature. It's something we want, but it it just goes to show again how perfect the UI is for Web Flow, how they've kind of nailed that, you know. Um Yeah. No one's really brought out anything groundbreaking. Web flow. Yeah. Yeah. That was composer, right? Yeah, that was composer. Yeah, true.

But web flow really nailed down and even though some folks still saying like uh framework looks a bit uh neater maybe it's like the contrast and the color and it's a bit simpler but still like web flow there is nothing that has come that is better uh than web flow like fundamentally better uh but again like this this I really liked it I really like having the idea of like maybe for you know if I'm doing caulk something like this would be nice with web flow or like clamp you know the the especially with caulk it can get really um long the syntax you have so yeah the only thing the only thing it doesn't have is obviously you can't drag and drop elements into this right you can't drag a form element you can't drag an input you can't drag a button so you have to first prompt the the UI get the UI or at least the framework in which you can then edit.

This is an editor, not like a creation tool, right? So that's the only other thing to to consider here. So again, just I wonder who this who is who who do you think this is for? Does this does this fit into your workflow? Does this fit into a an imaginary workflow that you don't quite use, but you can see that? Yeah, I can see this coming. So what I have here uh I I can see this useful. So um what I have here I'm building this as a prototype. Well the the UI here is a prototype. The main thing is this right the controls that I'm adding like the the essentially an animation engine for this type of animations is what I'm building like as a mini tool.

Uh and here like these buttons and the the HTML for these are generated with AI. So I'm asking AI, hey add that button and that paragraph and then it decided to give it let's say this border color. I actually didn't. This is like AI taste. Why should that paragraph look like a box whatever it does? So for something like this I could see it useful. So instead of asking uh AI to ah go change the style blah blah I'll just come here and let's say I I don't like the background I can just remove it you know or I don't like the border I just remove it done in instead of like fiddling with the prompt and like asking just it took me two clicks.

Yeah right in in some areas this could make sense. Mhm. Yeah, it's definitely I mean, you know, I was obviously being bombastic at the beginning here saying it's killed web flow. It's it's it doesn't replace like a creation design tool, but it's it's definitely a lot of Web Flow people were thinking about um you know, bouncing and going. They're all questioning whether cursor is the is the tool for them. And for a lot of lot of those people, it's not because they still want to they still want to create, they still want to design. But this will certainly make the transition a lot better that they can actually handle the thing that they've created cuz they you know again they don't want to write code or they don't know what to prompt you know it's it's nice little nice very welcome addition.

I just worry that cursor is getting very very bloated. Yeah, like man the UI I the the model uh let me see if I can uh show it like you know when I want to choose the model uh sometimes let's see if I do this sometimes this you see this window that pops up next here this little window it pops up here like on the left like it happens a lot of the times like all the times it happens and it happens with one of the models and not the other one like this is the tiniest thing like I want to read what it is and then I have to read on the other side like the UI is broken.

So yeah and even this editor and versus the agent view um or like the way you prompt it changes all the time. It makes me question like it why is it like this in this project? Why is it like that in another project? So, it is very bloated and I feel like a lot of this is because of it being very bloated. They just they're just adding and adding and adding features and it probably just needs a bit of a reset and like what what are people using? What can we get rid of? Um, yeah. And I think this is normal for an industry that is not figuring that hasn't figured out yet. Exactly. I I I feel like this is like I don't know like 200 what 15 or so no prior to web flow like when WordPress was still figuring out templates like it was so messy.

It is still so messy but after web flow figured out UI for code especially like UI for front code is what web flow actually like solidified and figured it out. M every single tool after is just a copy of web flow. It's just like naturally a a very sim like even this this is like a bad copy of web flow, right? Yeah. Yeah. And even they might not I don't know if they copied literally web flow or not but like the whole industry like decided to go this direction because it's a logical conclusion, right? It's a logical conclusion to land on. Exactly. That's what I'm saying. So, we still haven't figured out uh the the proper UI for AI in my opinion.

Like yes, there are like talks about like AI doesn't need UI. I still disagree. Like uh I still feel like we want to have control because we're such visual creatures. We want to to have that level of um that or like that form of information. There's there's an there's an interesting video because uh Open AI have just celebrated 10 years um which is crazy if you think about it. Um the opening thing they talk about the inception of chat GBT which is obviously prompt based input for AI where like that was the thing that was the thing that made people realize none of this would none of this is possible without chatbt because they realize what you can do with text input and they just sort of give a little brief discussion or he mentions like that's how they landed on it.

They didn't realize that text input was so intuitive and powerful compared to other forms of input that they had found. So it's it's it might not be ideal, but it's it's the one that makes the most sense. And I don't doubt that others have been explored. And in some ways this is, you know, is you you don't need to know that this is writing a prompt, but because you had that chat window open, like what would happen if I actually I know what happens cuz that's how I learned when you have that chat window closed. It opens up. But all you need to do is abstract that away and then you've got your your AI.

Be nice if they actually let's speculate. It be nice if they make this like open open s well they it is open source because it's cursor but like a way to add certain controls that you often use. Maybe it's drop shadow or maybe it's um I don't know maybe it's adding um tailwind and drop down to the things or whatever. But yeah um a code name Zulu has said he's enjoyed using uh or she I don't know whether boy or girl um boy by the looks of things um using VS Code. So uh although this is cursor, we we did a we did a showcase on uh pencil uh a few weeks ago which started out this this idea of like it won't take Figma long to either build a plug-in for cursor or VS code or it won't take cursor long to build this feature in.

So but pencil could be the the thing on VS Code, you know, it's a lot it could be your answer if you're into this sort of stuff. Yeah. Uh yeah, but at the end of the day, all of this is uh still in development, hasn't been figured out. So, Web Flow is adding AI and cursor being an AI tool is adding visual editor. That's that's basically two different directions moving towards the same thing. They are trying to own the whole vertical of creating for the web with AI and with visual input all in one place. H it's hard to know where it stops and where it begins, right? What what goes back to what we what you're talking about with a landing page.

What constitutes a landing page? Like what is the what is the product? What is the end thing we're trying to build? And at the end of the day, it's web websites and it's an app, right? So, um, the only thing I'll say, and one thing I've been really committed to doing over the last few weeks is I dedicate Thursdays to a code day. And I I'm trying very hard to continue to learn code because the things I learned 10 years ago still apply. I still know accessibility. I still know basic architecture of whatever, but technology moves forward and I want to still know how to code a website. So what I learned this week was testing and it's in preparation for a video that I've got coming out.

But it's it's learning how to write tests. And I went I go back I love learning fundamentals. I went right back to why tests learned, how you write them, the libraries you install and things like that. is staying on top of it so that I am the best AI coder so I'm able to uh shepherd and and understand what the AI is doing to know if it's doing a right or or good thing or a bad thing. So I'm I'm on a I'm on a mission now to get up to speed with some of the latest coding practices and translate that to my agentic coding arsenal. Do you know what I mean? Again, completely forgotten why I was telling you this.

That's the thing with old age. You just forget what you forget what you're talking about midway through about what you're talking about. But I used to be quite good at that. Um but yeah, uh staying on top of coding. Um do it. Do it. Uh okay, cool. Should we wrap up there? Yeah. Um, you you mentioned this uh Oh, I I see you have this Hank Green also video. Maybe we should just mention it for our audience. Yeah, we can do that. The water use. So, I I watched that video and I I enjoyed it and I was like, I'm going to send this to some and then I forgot. So, I'm glad that you watched it.

Here we go. Yeah. Yeah. So, Hank Hank Green, the the science sci-fi show uh cancer survivor um very funny guy released a video on the the the the white lies on this water usage of AI right and he basically explains we've spoken about it a couple of times to be honest but this argument that water use is um a a big issue in in the world of AI and the long and short of it is really that it's so it's such a hard thing is just as m your mic is just like Müller's. Who's Müller's? I don't know who who's Müller's. Um I'm I'm going to take that as a compliment. Uh um it's a it's a dynamic mic, so I can sing with this as well.

Um the the long and short of it is is that it's such a hard thing. Like the e the easiest way I can explain it is yes it uses water but like the work I'm doing would normally take me a month now takes me less than a week like so think about that like I'm not working as hard or I'm not traveling to work you might not be traveling to it's so interconnected all these different parts are it's hard to say that water usage up or down um that's basic know you watched it, right? Yeah, I watched it. But um yes. So, so the the whole thing I don't want to spoil it. That's why I'm like trying to think like how should I mention this good like you should go watch this video.

You will enjoy it. Not just learn something, you will enjoy it. I think he's just an amazing storyteller and he brings you through this like idea of how much uh water is being used by AI compared to other things that that's where it gets interesting because he compares it to other industries and where we are potentially a massive one right oh no was it cars what's the one where it's like the water used corn which we don't even eat yeah that's yeah Yeah. And also you you know for your garden uh maybe potentially. So there there is like there is um AI using a lot of water like there are articles on this and then you compare that to other industries that is also important.

Then there there is the note the the point that you make which is well it's making you faster so it is better um overall but also I would add that part that with AI we are generating a bunch of nonsense that we are using AI to filter out as well like there is also that aspects of it like with job and resumes people are using AI to uh to make like hundreds if not thousands of versions of their resume based on and then they send them in masses and then they um uh and then the companies use AI to filter them. So we are generating a bunch of content that we don't read. Like someone made a video I forgotten who it was.

Oh gosh, M's going crazy here. I just hear myself, but like super high pitch. Can you hear me? I just Okay, so I do need to I can hear you, but why you sound your voice is funny? Is it? Well, let me know in the comments. Is my voice funny? Do I sound like a chipmunk? Is it just me hearing it like this? Um I'm sure you can hear me. Even if I sound funny, I can you can still hear me surely. But um there is a video that someone released that saying you're using AI wrong. And I was like, okay, what are people using AI for then? And literally they're using AI to like do things you can do in a reax that you do not need an AI for at all.

Like this is obviously code, right? This is this code stuff. for just doing stuff that you can just filter with normal JavaScript functions. People are using coders are using AI for which just blew my mind. It would never even cons it wouldn't even cross my mind that that is something to do like people are are using AI for ridiculous things. Um there was something else that came came into my mind while I was saying that as well. But um it doesn't surprise me that some of these these people are are legitimately or illegitimately using AI. We're just lacking the ability to do any kind of reasonable um critical thinking because it's saying, "Oh, just AI will do it."

You know what's going on. I'm trying to see if I can fix your mic, but I am listening. Yes, I I I agree. Yeah. Yeah. Yeah. I don't know. The mic is You sound Let me know in the comments. There's there's plenty of you out there. If you're on YouTube, let us know. I don't know how this happened, but you sound like you just consumed a bunch of helium. Yeah, cuz I heard an echo and I was like, I can still kind of hear an echo to be honest. No, I can hear you like you've consumed helium, which is funny because like it should be like a filter or something. Yeah, it's not sped up.

Maybe it's bit rate or something. I don't know. But anyway, uh we we'll wrap it up there, should we? Might be might be a sign because we've been online for two and a half hours now. So, let's wrap it up. Um well guys, cheers for tuning in. That is all we've got time for. Make sure to follow us on Instagram, Commando Show. If you're watching on YouTube, give us a like, give us a subscribe. We're closing in on a thousand. And I'll go back to what I said at the beginning. Got any ideas for how we can build better content? Um, you know, provide you with more interesting things, stuff you want us to stop doing?

We're we're lovable, guys. We we're happy to listen to uh stuff that you really wish we didn't do or or stuff we should do, then uh yeah, let us know. Uh like, subscribe, all that good stuff. Join us next week for all the news in AI design and dev. I've been Sam and I've been Kavarza. Keep on vibing.