#66 Timothy Ricks Responds! GPT Images 2.5, Perplexity Hybrid Compute, AI Safety at Anthropic

My eyes deceive me. Are we Are we going live 5 minutes before? We We plan to go live. Are we I like how we said like we should go live We should go live early, but actually we ended up just being five minutes early. [laughter] We are ahead of the curve. That's what's happening. That is it. Yeah, we're not early. We're ahead of the curve. Yes. Hello, Instagram. That's what Apple is doing, by the way, with their watches. You know, they their watches always show 10. Oh, hang on. Apple is 9:42. 9:41 is for the phones. For the watches, it's always 10 because it's a smiley and it's like a nice place to be. Uh but they they they showed uh 10 past 9 past 10 instead of 10 past 10 because they are ahead of the curve just like us. [laughter] So Apple is like us. Uh and then for the iPhone they do 941. That's when Steve Jobs came to the uh you know came came like showing the first iPhone. He was like, "It's an internet explorer. It's a It's [laughter] a music player. Do you get it? These are not three different To be fair. To fair, you called that. You called that. Yeah. So that's that's about Well, Apple. I mean, speaking of which, well, speaking of which, I was going to say I was going to say I got some swag. You might have seen me around town. that. I want that. I really, really want that. I got Riplet, if you're watching, please. Rayar, if you're watching, man, [laughter] hook me up. I I don't want to wear this. I want that. Replet is cooler. Um, I also got an OpenAI tote bag and some stickers as well. So, but I've got them for my friends. So, um, sorry. Um, yeah, I went to Open AI. I didn't meet anyone. I don't think I met anyone from OpenAI to be honest, but I went uh I went to the event last night and got chatting. I saw Rhyar who's been on the show. I finally met him in the flesh. He was there. Awesome. [clears throat] We didn't get a photo though. I'm not I'm not really a photo person to be honest. I don't really saying that. You're a video person. Saying that, I just got a text about 10 minutes ago from a friend saying, "Is this you? If so, we were at the same event yesterday and it's it's a photo of me that Replet have put on their social medias, which I can't I think it's on Facebook. Um, that's cool. Uh, and that's a weird um notification I just got from YouTube. Someone's created a conversation from a comment that I made. I've never seen that notification before. Anyway, I digress. Uh so I have been I'm now in photos uh shared over replet and met met a few of the staff met a few people uh had a little bit too much to drink. Um so I don't know maybe I'll just and I've got a gig to go to tonight as well so I'm absolutely exhausted but it it will be fun. Um but yeah um replet some some cool dudes. Um and good to see Rhyma. So shout out to him. I think he's leaving today. But you asked, "Are they based in London?" No, they're based in San Francisco. They opened up an office in London. So that's where I was yesterday. And oh my god, I was such an idiot. Shows how much of a charmer I am, though. So I Googled. So I I responded to the RSVP that Replet put out to say that we're opening up in London. If you want to come to the party, RSVP. So I RSVPd. And I was a bit anxious about not getting in. So I messaged our friend Rhymar who works for Replet and I was like, "Can you get me in?" He says, "Yes, if you arrive early, I'll be able to get you in." So time rolls around, I get on the train, I Google, sorry, I Google Replet on the on the phone, find out it's in King's Cross, get on the train, head on up, have a little drink cuz I'm there really early. I'm there at like 5 and this thing doesn't start until 6:00. are on there like I'll have a I'll have a cheeky half and I'm like I'm not getting a response from Rhyr to be fair. He's very very busy. So, you know, but response from but I'm hanging around the replet offices and I I have two halves so I'm starting to feel a bit you know loosey goosey. So I go to the offices and I um ask reception where's Replet? They point me to Replet. Sure enough, there's an event going on, but it's a little bit quiet. So, I explained at the explained at the front door of Replet and I said, "Um, I was invited to this by my friend Rhyar. Uh, I don't think I'm on the list or anything." And so, they wrote me down and took note that I was going in. And I walk in and then they guide me through into this little room that's very quiet and there's four panelists up on the stage and they're talking about data bricks and replet and the partnership and the the chief revenue officer is on the stage and it's all very you know it's all very um how do you say suits and and businesslike and I'm rocking up in my hat. I don't I'm not wearing the t-shirt, but I've got like a Marge Simpson on my t-shirt just looking all kind of sexy Marge Simpson looking like a complete teenager and everyone around me is in suits. [laughter] And I'm like, okay, whatever. It might just I might just be mega early or whatever. Some sort of ism. I'm I'm continuing to have a few drinks because they got free drinks there as well and I'm feeling a bit whatever. And I actually raise my hand for a question and I talk about I ask a question to the CRO of Replet and data bricks and whatever and I say you know I think it's a good question and I say you know we're in a world where we've got a bunch of engineers who have been writing code for the last 20 years who are really good at using a agents and AI. How do you uphold that that um as the bar is now very very low? How do you maintain quality with people who don't have that level of experience, who don't know? And it created a big discussion. People came into it. So I was like, I'm like a little bit drunk, but I'm like talking to like this. Anyway, [snorts] I walk out um to get a drink before the next lot of drinking in London. Lot of drinking in London. [snorts] Uh go out to get another drink before the next talk. And someone shouts over to me and says, "Nice question." So I'll like walk over to him. I have a little conversation with him. I was like, "Yeah, I thought there was supposed to be like a party or something." He was like, "Oh, yeah, there's another event going on somewhere else, but I don't know where it is." I'm like, "Uh, okay, that makes sense." So, I then speak to one of the replet staff who was who handed me the the, you know, lanyard and she was like, "Yeah, it's uh it's at South Bank Center or it's down South Bank." I'm like, "Fuck sake, I've come to the wrong bloody wrong event." So I hop on a train. I rush down there and I get there and Rhyar was there sorting people out at the door of this building. He was like, "I told you," he said, "I told you arrive early." And I was like, "I freaking did, but you'll never guess what happened." And, you know, briefly explained to him. He got me in. And then I went up and then the CEO was talking. Amjad, I think his name is, uh, was talking and and Open AI were there as well. And, uh, yeah, it was just drinking and eating and getting to know people like a like a networking event. But yeah, such an idiot. I don't know. It's because I wasn't actually accepted to the event. If I was accepted to the event, I would have actually probably seen in my emails like where the event was, but I completely had my assumptions. Um, if you had like Apple intelligence, it would tell you because it context. Well, there's a nice segue to because I have nothing more to say about the event. It was very fun and I've got some swag, but I'm actually Yeah, I would love to join if like it would be possible. I love that kind of stuff. Yeah, I um my friend really wants to work for Replet. She's a designer. That's the person I got the swag for, but it was a bit cheeky kind of getting Rhyr to sneak us in and and whatever. So, uh, yeah, but, um, yeah, maybe they'll do another one. I don't know. But I would I would work at Rep work at Replet. I would I would work there. I I think they're like a fun company. Like Yeah. Yeah. And they are ahead of the curve. We've been talking about this. [laughter] We're so ahead of the curve. All of us. We're just so ahead of the curve. Yeah. Yeah. It's all about being ahead of the curve. So, we It's so weird. We've got We've got a couple people from my Twitter. We've got more people from YouTube than on from Twitter, which is It's some It's amazing. Often the other way around. I mean, it's amazing. Yeah. If you're on YouTube, take a moment to give us a little like and a little subscribe because I sent a tweet earlier. I want to get to 5,000 subs by the end of the year and uh I think we can do it. We just need your help, dear viewer. We need your help. Uh but yeah, we went live earlier this week for the Apple Rise and Shine event. We're not going to go over in depth all of the things, but um quite cool. I'm really excited for the Fold. I will be probably be getting it under a low payment plan thing because, you know, I don't have two grand lying around for a freaking phone that I don't even know if I'm going to be using. But there you go. Um, and and we saw it the opening animation. I think it's on this page. I think I will show you. Yeah, I have a bunch of like cool stuff I want to show the the audience. Well, apparently live at the event that was just everyone just jaw dropped at that and it's such an awful thing, isn't it? Just unnecessary. But I came to the page. I saw the page and the first thing was like hello hello and I was like this is so Apple and I love their copyrightiting man they are just so good you know Apple duo and hello hello hello it's just just so good it's cheeky but I love it but the most important one as you said is this animation that I want to show you here. So you click on this foldable design if they show it sometimes they don't show it because it it sometimes it doesn't load it because it think I'm not on a good internet connection. So, it loads it. See, now it's 3D. The first time it wasn't. So, look at that. And I saw that the I saw that the the the thumb has a little bit of resistance to it as well. The thing that you're sliding has a little bit of resistance to it, which is interesting. Look, it it's sticky here. You see? Yeah. It's just like little details. It's sticky. And this is just so beautiful. And obviously people see this and they make things like this. This is just I love Yes, I saw this. Yeah, love the internet. I know it's it's not useful or anything. It's just like How are they doing that? Because I don't think there's a sensor on this. Yes, there is. Yes, there is. Uh the Yes, the hinge has a has a sensor. I have actually installed it. There there's a g repo you can install. That's crazy. That puts a door knob like a a door uh not knob the door hinge sound like you know the quque the quque is it called Yeah. [laughter] It puts it on your Mac. So when you close it it's like which is just so funny. That's funny. What an unnecessary sensor. Like I understand opening it and closing it but the exact angle. Yeah. That just feels so unnecessary. So you actually have that. I have that. You also have one. I don't know what they use, but you can tap your Mac and it can recognize it. So you can use it. Somebody used it for spanking. So it says spank it on me. [laughter] Yeah. And and you for the touchpad actually you can um you can use it as a scale. It's it's quite good. Like if you don't have a scale, you can use it as a scale. It knows like up to five grams or something or I think it's like even less than that. So I I did try it and it was pretty accurate. Anyway, people are also going crazy about this. Something that I at the event in the live I said ah I'm not sure about it. I don't really love it because to me it's not very familiar like this is not a battery and this is not like you know and also what if you have two SIM cards. So there are like UI talks that I thought like are relevant to our audience that I wanted to bring in. Um, yeah. Uh, like really cool stuff. Here we see a 3JS model. They've been perfecting this. They've been perfecting this. Actually, there have been people in the community in the 3JS community. Uh, every time they overanalyze what Apple does here and they they check how Apple does it like so perfectly to be very uh like high quality, high fidelity, but at the same time lightweight. So, it's like the perfect balance. And yeah, people who are doing award-winning websites, what do you think they do for other companies? They do this type of thing. So, uh, to all the people who say like these award websites are like they have no benefits. They can lend you a job at Apple. They can land you a job at Apple, but they can also sell the heck out of a foldable iPhone. So, yeah. So since we did the the live we watched the live event and we shared well we didn't share the live but our experience I it's it's been growing on on me and it it's quite expensive but if you want if you want me to help you with a project if you're a client and you want help with a project maybe maybe I I can help you if you can help me with this project. So, uh, yeah, I I I might get it actually. It's really nice having that having that screen that Yeah. opens up. It's just it think fast. What do you think you'll do with it? Why do you think you need that big screen? Uh, Tik Tok and Instagram scrolling at the No, I'm kidding. [laughter] Um, honestly like I feel like the the do the Duo app is really nice. The media consumption is nice, but most importantly I want to get like to see if I can get some real work done here. Yeah, that's that's for me it's the productivity. Yeah, that that's really nice. And I can probably mirror my Mac window. Figure out a way to mirror it to here and have my Mac window all the time with me. Yes, it's small and like it's not the same thing, but it's it's it's close in aspect ratio. It's not the same, but it's close. But it's really useful if I can get that done. I can actually, you know, let agents to do the work, but I can do proper reviews, not a phone review, a desktop review, which which is really useful. Um, it might even be something like a monitor like I, you know, to to see what your camera is seeing as well, you know, and I wonder if that, but I would probably do that on my iPad if I'd if I needed it, if I felt like I needed that sort of feature, but, you know, again, just the versatility. I think it's a combination of watching I do watch a lot of YouTube on my iPhone, which is just crazy in bed and stuff. I I sometimes can't bother to get my iPad out and it's just not as the tent mode even. Do you know what I mean? That that would be cool. Um, and then as you say, the productivity. I was tempted by the other one, the fold, uh, the the Galaxy Fold or whatever they call it, just to see if I would do, you know, it could be a clue as to whether I would actually use the Duo. Might be a cheaper way to do it. I bet there's a bunch on sale now or whatever. Um, but it's nothing about nothing like being baked into the Apple ecosystem. But I don't know. I don't know. And the camera I said the camera is worse, which is true by the way. Like the camera is not as good. The only camera that is as good is the uh wide angle camera, which is not as useful, but the main camera you could use it instead of the front camera. You can open it and use that. So for me doing like sometime from time to time, you know, social media type of thing, I think that would be also handy. Like I can maybe I can just use the main camera for both cases. That would be really cool. Something else that they showed and I think a lot of people missed it. I don't know if I can find it here. So, you can you can have a video chat with somebody. Yeah. And you can share the front and the back camera at the same time with Yeah, I think I think they showed that. But like imagine the battery power that needs. Yeah, but like there is no reason to limit that to just this iPhone. I I want this to regardless of this iPhone to to have it. Yeah, that's true. Actually, it's only the screen really that you've Yeah, that's a bonus there. Everything else can be achieved. I want to Yeah, I want to have that capability. Anyway, also the website is broken for me. Like you see this green? This this is not uh this doesn't have enough contrast and it's been like that. Um maybe there's an extension or something of me breaking it. Could be. Could be. But uh anyway, um and and yeah, where is iOS 27? I thought that normally gets released as well. And they've got to release it before maybe it's the watch that's getting released the soonest. Or they've got to release it before they start shipping the phone. They can't ship the phone with iOS 26, could they? So, where is this iOS 27? Uh, yeah. I don't know what's happening with that. Interesting. Anyway, let's We started early. Let's Let's see if we can, you know, wrap this up. Keep the momentum. Keep the momentum going. Keep the momentum going. We didn't go with an intro, by the way, which is Well, I thought this was more of a chat than a than a than a highlight article, but let let's let's go through Yeah, go on. Let's do the intro now and um we'll jump into Perplexity. All right. So, this week we had Apple event which we just talked about. We have something really cool from Perplexity called hybrid compute. Uh we have Timothy Rick uh responding to him joining web flow and we have like a lot of things to talk about that uh with source web flow run segal and flux academy and that kind of stuff. Uh we have open AAI chat GPT image gen 2.5 which supposedly is faster and better and I have examples to show you. We have Lovable launching a partner pro program. Tailwind just joined Shopify. I never saw this coming. And an AI safety researcher from Anthropic quit because he thinks uh we are all doomed and we are gonna die within the next four hours. [laughter] Was it was it two? Well, that's what he thinks. We will go into the tweet and read it. And he actually he seems to be very serious about it. All of this in today's very serious episode. This is Command AI. I'm Kavarza and I am Sam and this is a week in AI design and dev with a serious face. Yeah, very serious. Very serious. Uh we have serious stuff. Oh, and there is one more thing that I will add and we'll talk about it. M surprise surprise. Well, I got to show you this really cool feature. And before we get into it, this is uh we're proud to be partnering with Perplexity for this feature. So, I will go ahead and check this out that I prepared. Boom. Look at that. Oh, it's messed up the the wait the it's messed up the uh Instagram thing. Wait. Um what did you Ah, yeah. This perplexity thing at the bottom left. Look at that. We're professionals. How can I now? Now, this is the That's okay. This is This is the one for uh for Instagram, but I'll be sharing my screen now. So, basically, as token uh costs increase and agentic max like this baby uh becoming the norm, more and more people, I think, are looking into local AI to reduce the costs. Um, but the big issue is that AI just isn't there when compared to like the frontier stuff, chatbt and anthropic. Well, what if you could just have both privacy and cost savings of local AI and the raw power of code? And that's where Perplexity's hybrid compute steps in. Honestly, again, sponsor segment, but we probably would have covered this anyway. It's pretty damn cool. And I would cover it only for the website for the least. It's it's just so I mean they're they're designers. We know they are just I love their brand. I do. I do. I love their brand. Um let me find my tab. So to give you this in a nutshell um it's first of all the app is great. The the desktop Mac app is amazing and low key it's amazing at computer use. This is not what this is about. This is about hybrid compute. But the computer use thing is by far and away so much better than Claude. Most things are better than Claude. It doesn't take much to be better than Claude, but this is much better than Claude. Uh I sent something yesterday to try and send me some information from a text I'd received on my computer or whatever. Claude just fell at the first hurdle. So I just sent it to to Perplexity. It got it sent it back. I'd strongly recommend the uh for it for that alone. Um, but you see here they've got this uh Mac Mini and I think they're kind of pushing this to the same sort of people who use Hermes, who sort of use Open Claw and that kind of thing where they've got that dedicated machine set up to run aic task to always be on, always available. And what they've added is this thing called hybrid compute. Basically what it does is when the when perplexity receives a request and you have hybrid compute turned on, it delegates the private sensitive uh information such as name, addresses, uh just generally sensitive things. It gives that to the local model. So it's all kept on device and whatever. and then everything else the planning and the heavyduty stuff it's given to the the remote models um or the the the cloud models. So I really like I think that's perfect. I think that's exactly what we need at this point whilst local models tend to be not as you know complex but they they really it's undeniable to have private local data running on local AI models. Do you know what I mean? And then the rest that's not whatever. So, with that, it's it's kind of um let's scroll down here. There's basically and I'll go over the security stuff. Don't know why it's so low down. Um it's designed to run on Apple Silicon. So, sorry Linux guys and Windows guys. It is an Apple Silicon thing for for the moment. I don't know if they have plans to do whatever, but it's designed to run on 24 GB uh unified memory max minimum. 32 gig is for the best results. And that is because they give you a choice of three models to pick from. And in fact, have you got why don't you we should have planned this. Why don't you download Perplexity? Because their intros are so nice. Their intros are so nice. But they basically give you a choice of what model do you want to process on your local machine. And these are the three models that they they they just pick for you, you know. Um, no alarm. Get that done fast. Yeah, maybe. Maybe. Let's see what you can do in the background there. So, no alarm or anything like that. And And it just works. It's literally I sent it a task yesterday to um look at the we've got a contract. And look at this contract. Compare it against our insurance policy document. So, plenty of juicy private information there. I just sent it to to my Mac or I sent it to Perplexi app and sure enough, let's find it because there's a lovely little um thing here. We wait around for this. It brings up this cool little thing to let you know that it's running a local task and how many tokens uh it it's saved you basically. And then on the left hand side here, it pops up with how much it cost you in credits that you you buy from Perplexity and how many tokens it was used locally. And I these two I did two tasks. I did one where it was just remote. So yes, I sent all my private data to the cloud. So you know, whatever. And then one hybrid. And the one that was cloud only took about 2 and a half thousand perplexity credits. It was using Astra. But the point is the local version, the same task. The local version took 700 credits. So that's like a third of the cost running it uh on a hybrid solution. So it's a really good cost-saving thing. And of course the privacy thing. So yeah. And then the only last thing to say, it's got this PII, this personally personal personally identifiable information, which is another model that you can download through it that really just determines the sensitive information, names and things like that, which is again all run locally. So really, really quite nice. And like I say, this comes in clutch when you stick it on a Mac Mini um and you just have it running all of the time sort of thing. Obviously, [snorts] local AI is a little bit slower. I wouldn't be recommending this for tasks that you just need, you know, that instant access, but it's still if you can schedule things, if you can have it run in the background, if it's non-urgent, it's a really good solution. And you don't always have to just do the hybrid thing. You can just send it to the cloud and and whatever, but um yeah, you can, you know, it's it's not you're not locked into one or the other sort of thing. So very very cool feature from Perplexity. Uh like I say, they partnered with us to bring you this section, but we would have absolutely done this anyway. [snorts] Did you did you install the app? Yeah, I want to show you. So uh it was quite smart to also pause. I don't know if it's Proplexity or just Mac. Um I don't have this u window uh active anymore. So, it paused. Uh, now it's continuing. I don't know if you can hear it or not. No, but this is lovely ethereal music, isn't it? Yeah, it's it's just so good. So, this is the shortcut to for triggering perplexity again. Nothing to do with hybrid compute. It's just perplexity. Oh. Oh, is this you? Oh, this is chat GBT. That animation. Why Why did it work like that? I don't know. Yeah. Wait. Uh let's see. It's uh it's probably taking Yeah, Chad GPT has this already probably. How can I get rid of that? Oh, this is the other things that we plan to show. Um there it is. Oh, it keeps disappearing. Is it minimize all your windows except for Oh, there we go. Yeah. So, yeah, the double command brings up continue here. Yeah. So, the double command that you press brings up the what do you call it? Spotlight thing for using Plexity really quickly. You can drag files in all of that good stuff. This is just the Plexity app. Yeah, it's really nice. Um, their intro is really nice and I can hear it also like um holding the function key. Beautiful. Oh, you don't you don't So, I don't need uh Whisper Flow anymore. Well, if I use that Whisper Flow is good for writing your emails, but if you want to trigger your agent, then yeah, you can choose that. If it can copy it uh to my clipboard, it'd be amazing, wouldn't it? Yeah. So, okay. Now, don't dox yourself. Yeah. But yeah, really cool intro. I'll do that so you don't dock yourself. Yeah. Um, and the only unfortunate thing is obviously it needs to download the model for you to see the actual user interface, but maybe I can share my screen um to show you what that looks like. Maybe maybe that we'll we'll do it up to we'll do it up to your uh what do you call it? Your uh downloading the models and then I'll take over from there. Uh okay. I think I did have a pro account at some point. Trying to Oh, we should I maybe I should cover pricing actually since uh since you're talking about that. So, pricing is like this. You've got basically the $20 a month which gives you access just access to computer use. Right now, they're giving you 4,000 bonus credits, which is quite nice. But you will have to top up your credits. I'm not sure that looks very good on um what do you call it on on Instagram? No, [laughter] whatever. Instagram can't see us. Okay. So, yeah, you uh you will have to top up your credits and that will give you sort of that added flexibility. But if you go all out and go perplexity max, you get reoccurring 10,000 monthly credits. So, um, definitely worth considering the local hybrid compute if you've if if you've got a powerful enough Mac to to actually do that because it will save you a lot as I experienced to go with that with the hybrid solution. So, yeah, that's that's the pricing. Nice. Uh I'm trying to remove that double command uh from chat GPT. Oh. Oh, you've got double command there as well, is it? Yeah, I I didn't I have not added that. I think this is their default or something and I can't find it. It's uh I'm looking command command there are a lot of command stuff but double command maybe I just search the screenshot because it if I do double command it screenshots and puts it into GP. Okay, no worries. Why should you do that? I'm going to show what it will look like um when you've downloaded those models and we'll we'll just Yeah. add them together. So basically um under so this is what it looks like. This is the Plexi app. This is what it looks like. And under this you can obviously select whatever model you have whatever model you want. And you've got things like I can see there Kimmy K3 Gro 4.6 Gulm 5.3 Deepseek even and then obviously Fable Opus Sonet uh GPT Soul GT GT6 basically. Um but then you got this hybrid toggle thing. Sorry for those. I can't really maybe I can actually. So I can increase it a little bit. you got this hybrid solution and I've got the perplexity model there for you know the the local model and then when it needs to send it off I've got comm3 and whatever I if you we click computer there whatever I type here will do whatever I ask it whether that is a computer use task go off and get a file on my computer and make it add this information and make it into whatever it is if it's computer use or not it will divvy up those tasks whether they're sensitive information versus whether it needs all the compute that uh the Frontier models, Frontier Labs uh provide. And that's kind of it. And it's pretty cool. Yeah, that's cool. Uh regarding the the thing that you want we wanted to show just sharing my screen uh is probably easier. I think I need to upgrade to Oh, okay. Right. Okay. because I can. Yes, that would make sense. Is it local in it? Yeah. Yeah, that's it. You can see it's all grayed out there. So, you can select a model. Yeah. There's just a nice little interface that takes you through downloading the local model and setting up hybrid. But, of course, you need Perplexity Pro to have access to computer, which this is what it all depends on. So, no worries. It's all right. We'll we'll wrap it up there then. But overall, really really cool feature. a really nice feature. And I'm a suck. I I I do We'll take this off because uh these these are genuine thoughts. Like I do like I use Comet as my browser. I do really genuinely really like Perplexity and and their design and the attention to detail is really really nice. Um so yeah, cheers for Perplexity for for partnering with us. Yeah. Cool. Wow. They got their money's worth there. We really dug into that, didn't we? That wasn't supposed to be that long. [laughter] Yeah. So, people are already asking, "Did you guys intentionally skip over Tim Rick's news? I didn't catch your commentary and opinion." That's coming up. That's coming up. No, we haven't uh share this video with the community. Let's [laughter] Let's have an uh live conversation about that. Yeah, let's do it. Yeah, let's do it. So, we we'll we'll get into that uh in a second. Cool. Wow. No, we'll do it now, right? Yeah. But after answering this question, any chance you guys are thinking of building a website builder someday, maybe an astro specific one? Not really. I like Sam you can speak for yourself but in my case I'm building actually some I I don't know if I should call it tooling but at the end of the day a lot of it is either just Nex.js project uh projects that I'm spinning up and also like MD files around it. I'm trying to create some sort of like semi-automated ways to create websites that are like I'm not talking about like award-winning websites or like the typical big 20k plus projects that we do but actually smaller ones but in a very contained and predictable way with AI with the speed of AI but with all the controls on CMS and everything else. I I I might one day uh open source it but I'm getting close. I'm getting close at like I think a lot of it is figuring out at this world and in this world that we can do anything that we want in a way like if we want to we could try to get something like to create something similar to source right um given enough tokens and you know a few engineers but like it's easier than ever I think I'm just trying to wrap my head around the thing with the direction that I want to create that for myself and then I will I will share more about it and I will share it probably with the community as well. But if you ask myself, I think I'm on to something. [laughter] I think I think I'm on to something. There's some creating some like really nice tooling around simple websites for small to medium-siz businesses. essentially in a way my own builder. Yeah. With AI. Yeah. Yeah. Um my answer to the question, it's a really good question actually and I didn't I was outright going to say no. Don't be silly. Then it did get me thinking, but ultimately like I don't I don't need a web I don't need a website builder, right? Yeah. I'm very h and we'll get into it in the in the discussion with Timothy and and whatever, but uh I'm very happy with using Claude. I don't need a drag and drop interface. You are happy in the terminal. I'm happy in the terminal. If I need to give my clients access, there are already so many great tools out there. Ship Studio, um, you know, maybe Source when it comes out, Timothy Rick's, uh, Stacky, there's so many great tools out there that I would I would be hardpressed to say, well, what can I bring to the table? So, I'd fire a question back to you, Voltz. What would you what do you think people like us can bring to the table with regards to all that? Uh is there something you feel like we have a specific edge to to be able to do something that all of these great products can't do? Because it's a lot of energy, isn't it really to build a tool to maintain it and all the rest of it especially when it's not a tool for you for yourself. I have everything I need. I've I've you know I've got an ecosystem of apps and things like that. I may end up building let's say a CMS to my liking or whatever but Convex is great you know I may end up building a database hosting solution but uh again convex is great point is uh until a client really needs it I probably won't have the time to just build something. So yeah, I'd love to hear what you think we could bring to the table. Uh but no plans to do anything like that yet. Web flow or wow, Timothy Ricks has responded to us. Maybe maybe um there's two brilliant videos that we will we we think you should be watching right now if you're at all in the in the let's say the what do you call it community ran seagal's web design is changing and Timothy Rick's I accepted a job at web flow because if you didn't know he accepted a job at web flow and there was a lot of speculation we speculated on the show which uh Timothy actually responded to So, we know he's watched our video on it. Uh, and it it clarified a few things, but it also, you [snorts] know, it left a few doors open for speculation, which we'll we'll cover now. So, so definitely stay tuned for it. But a few things he covers in this video is he's closed down his community, and that was not something that Web Flow asked him to do. He noticed that a lot of the designers went to Framer. A lot of the developers were going to AI to answer their questions a lot faster than he could get around to doing inside of his community. So, it wasn't it his his audience, his I don't know how many members there were in the community, but it was a dwindling amount of people. And once he got the offer from Web Flow, he he was like, you know, and that wasn't a big source of income from him then and he's got a family and all the rest of it. So when he got the job from Web Flow, that was when he decided it to to close it down. Again, he really stresses a couple times Web Flow didn't ask him to do anything. So in terms of his projects, Lumos uh for web flow is going into maintenance only mode, especially as source come it might change. I I this is me speculating. This is not what he said in the video. Source might change his mind on that once he sees how it all integrates or something. But anyway, so Lumos for Web Flow is going into maintenance mode. Lumos for Astro he's still going to be working on and he believes that will definitely work very well with Source. So that's cool. stacky, he's basically leaving it, he's giving it to us as an as the open source community to contribute to to to bug fix and all the rest of it. So, if you want to get involved in that, you want to get your name on some open source projects that really help the community. Again, uh who was it that asked the question whether we build our own tool, get, you know, if you want something, just tweak stacky to your needs. Modem, which I didn't know about. I'm guessing it's some sort of computer computer uh component library. Uh that's that's he's left it up because he thinks it might be helpful, but he's unsure of the future of that. Um I guess it's basically in the same maintenance mode as Lumos for Web Flow is. [snorts] Um and then he goes on to his thoughts and feelings about Source, which he hasn't had access to yet. He wanted to make this video before uh he had access to Source so he could come up with some speculations and thoughts and feelings. And he he thinks it's a bold move. And we'll get into Ran's video about all this. He thinks it's a bold move to release such a overhaul of a platform that's really well beloved by a lot of people, right? It is beloved by a lot of people. So, a lot of designers and whatever. So, um but he did he also mentioned the content he will be producing will be astro focused. So, it'll be teaching developers how to step into the role of someone who can use Astro. So by the looks of things, he's not going to be producing web flow content, which is really interesting. Um, and his role at Web Flow is a forward deployed engineer and a small community-f facing role. So really interesting. He thinks he thinks source will open up to a new audience of Webflow users and um he enjoys the flexibility that it gives uh people who build the they can give their clients a platform to manage the website that they build in Astro and things like that and allows them to add a CMS or a flexible CMS depending on uh well obviously you're not bound by the web flow CMS anymore. So, um, and he really doesn't want to see the the no code option falter in Web Flow because of course they believe everyone's moving to agents, but he and we said this as well, still want to be able to see this drag and drop UI, you know, still wants to be he wants to see that it's you're you're still able to build inside of Web Flow um, regardless of what they do for free, drag and drop builder and all the rest of So that's what he said. Uh a few things I take from that. I mean do you what have any thoughts on that rather than me just waffling on for too long? It's no we will get into like thoughts but well it it sounds all good like yeah what we expected sure to be. Yeah, we had like, you know, you our thoughts, but yeah, it does make sense everything that he said. So, and the title of our video that he responded to was like, um, he said he's not trying to kill Web Flow. And we we never thought he was go he's trying to kill Web Flow. This is before he received the offer, by the way. [snorts] Uh, we never said he was trying intentionally. It's just by um uh proxy, by accident, call it, you know. Could he? Has he accidentally killed Web Flow? It's speculation. It's just a fun title. We never thought he was intentionally against or trying to kill Web Flow. This That's ridiculous. Yeah. But my my take Oh, go on. No, no. Go, go, go. So my take from this video really and and it really brought back to the video that we made and something you said which was it you don't trust web flow and whilst he emphasized they never asked him to you know take anything down stop producing anything do this do that [snorts] I would be very hardressed to imagine that it didn't cross Web Flow's mind that if they took on Timothy Ricks that they have removed a potential threat because he's so influential in the web flow community. If Timothy Ricks jumps, so does the web flow community. So to that I would say that's my only thought. If if we are a bit if we are like you know [clears throat] thinking okay web flow is trying to plan this maybe but he has already jumped so I would maybe give maybe I would give web flow the benefit of the doubt actually maybe saying that if web flow is killing web flow if web flow is killing web flow and source is the future so maybe having one of if not the biggest you know uh community leader in the space on their side is a good idea having him build with Astro but imported into source is in their best interest right so maybe it is also an investment web flow like seeing okay in the future there won't be the web flow you know designer won't be the main focus it will be all source so why not have this by um bringing whatever he builds into that platform and showing clients what's what's possible because I think this is what Timothy always has been really great at. I think he has done in my opinion two like so many builds that he has done in the past. They were like things that I probably would not do because they were just so niche things, so geeky things, but they were amazing. They were like, "Oh, you can actually do this with grids without using like break points." There are so many things that he's like just inspirational. He's like showing the community, showing potentially the clients what's possible. And I think that's maybe why Web Flow one of the reasons they stash them. 100% absolutely 100%. And and I know our video will get watched by a bunch of people and I know we are quite let's call it an influence on the Web Flow community. So I'm just going to say it here. I'm saying I'd be hardressed if it didn't cross their mind. It wasn't the reason why they hired him. He's a great developer. He's a great asset and they are very lucky to have such a great guy working for them. It was absolutely that was the reason. But they they they got scared a little bit and I think that probably just, you know, we what what can we do about this? And then it's like, yeah, but ultimately they're going to be paying him probably good money to to to do what he does best. My business brain, this is a this is again, this is my side thing. My business brain tells me that I'm uncertain of the future. Was it a B was it a good move for him career-wise to go to Web Flow? I'm not sure. I'm just because I'm uncertain of the future of Web Flow right now. I don't I don't think they met the bar that was expecting and we'll get into Ran's video really quickly uh in a bit. My specific feelings on that. If I was a mate of his, I would have been like, "Are you sure you want to be joining a company when their future is so uncertain?" But yeah, he's got six months probation. He can hand in his notes at any time. He's feeding his wife. He's feeding his family. You cannot fault the guy for taking a a job offer from a company that he's loved for so long. So yeah. Yeah. You know, I I fully agree. Even web flow being a I don't want to call the shipping sync, but I'm saying this for the lack of sinking ship, buddy. What did I say? Shipping sink. [laughter] that shipping a sink across the country because I was thinking of a better word because that's not true. I don't think they are a sinking ship. I don't think so. But like accompanied at at least in the community like the community is not super happy. The community is like a bit heartbroken. We see this in our comments by the way to the on the video on Webflow about Webflow Conf. People are literally saying they are heartbroken. [snorts] they they don't feel like Webflow is talking to them anymore because Webflow is talking to marketers. We are not marketers, right? Anyway, we're builders. Um yeah, so to your point to your point, Timothy, he joined because it's still even even if they are moving away from the company, it's the it's the company that we've been engaging with for so long. Like I feel like I have friends at Web Flow because we've been like we meet all the time and let's not forget like we meet we met originally through web flow and I have like so many friends because of web flow. So we're absolutely we see that side of it as well. Um this is a nice uh comment and again testament to the um the um what's the word I'm looking for? the asset that Timothy Ricks is. He showed us that one man can build a great tool for the people because he's so embedded in the community as well. Let's not forget that he's so close to he knows what the community want and and that's why at the end of his video he said he doesn't want to see that falter from Web Flow, their focus on the drag and drop builder, the the way that we that design yeah designers like to use and inter interact with the website. So, [snorts] um, just reading this comment on the fly. Also, Timothy just listed personal reasons that family. Yep, we said that. Yeah, absolutely. He's, uh, he's so right to to be able to and so privileged to be able to accept that offer and feed his family whether it's the right move in 6 months or not. Point is, he's feeding his family and he's not trying to like we're painting it like he's some casualty or some charity gauge. Absolutely not. Like, of course he's he's probably got money, you know? Yeah. Yeah. Absolutely. um he's he's probably getting like paid really well and and I and I think he's not getting paid well just because of his community and his reach. Just to make that clear, I think there are not many people on the planet earth like literally like you probably can count them on your fingers. people who understand CSS, JavaScript and web flow to that depth like the web flow as a platform like how does web flow work versus how CSS works and how because there's a translation layer and he understands that probably better than anyone else. Yeah. So, it's a no-brainer to hire him and the community is just, you know, that that he has is maybe just in addition to that. Yeah. Um, I was going to move on to not totally move on to Ran's video. Did you want to talk about this comment before I do that or after we're done with Timothy? Let's talk about it uh right now. So, we have this comment. I actually have it here. Okay. [snorts] So sharing my screen and actually let me zoom in a lot so we see this better. Where did it go? Okay. Uh they most definitely needed uh Timothy Ricks to keep the designer relevant if source doesn't get the adoption they think it will. They need a plan B. And I I kind [laughter] of agree with this. Like it does make sense to me that why is Timothy is not working on source? Why is he working on the current designer? That's because he's an expert there. And if source doesn't get the adoption or as you know fast as web flow thinks they still want to keep the designer you know something that they can continuously work on and like push updates and who's better than Timothy and maybe that's a plan B as well. And this comment was also interesting like two other points. The whole marketer plus AI is as good as a developer AI needs to stop. And I kind of agree on that as well. Like no hate on marketers, but I I don't think like a marketer with AI is necessarily as good as a developer with AI. Like these are different things. Just just to make it clear like a marketer with AI can probably make a better converting landing page but a developer can create a better dashboard a better CRM like we are building CRM left and right for you know custom stuff for clients. So um I don't think like everybody can do this. It's something that you know you're focused in one area not on every area. So well just on that real quick is that you've compared building websites which a marketer would do to building CRM which you which is a dashboard which is totally internal which is a web app which marketers wouldn't touch with a barge pole. So from a from a perspective of building a marketing website, I I don't know. I think marketers are well equipped to build a decent marketing website like it's so it you know no disrespect to anyone building marketing websites but they are they're very easy especially for AI to do. There's no complexity in just hooking up a CMS having a page display the data you know the effort comes in to the copy to the UX to the design. So that's that's all I'll say there for for a baseline for a simple thing. Yes. But when you are talking about you know an Apple an Apple website [snorts] uh page like you know for their product that is that a marketing page I would argue so yes it is a marketing page. Yeah. Yeah. Can a marketer build that? Hell no. Like no way. But they're going to Yeah, they're going to they're going to But this is talking about or potentially like marketers 80% of marketers will be able to do because very few websites need the that Apple level you know interaction but I think it's becoming there are exceptions. I agree. I agree that there are exceptions but as we will talk about with the video from run the brand really matters and as the baseline as like everyone is pushing towards like this average it's more important than ever to stand out and to do that you probably want to do things even more crazy than Apple. it uh it doesn't mean to be 3D uh just to be clear on that but there are just things that you can do with like very simple lightweight shaders that I don't think a marketer necessarily knows about and these are not I'm not talking about the technicality of shaders I'm talking about the brand that you can nurture with like these type of like custom shaders and image pipelines and essentially like brand assets right so my point being a developer with AI is always more technical than a marketer with AI and I don't think anyone would disagree with that and it's just a point that web flow shouldn't present themsself as a marketer only platform it is it has been a platform for builders and [snorts] I find it quite sad that they are moving away from this builder story and they are moving towards marketers which doesn't speak to me at all doesn't speak to any of my clients like it's just not the space that I operate in. And I do understand they exist, but it's just not where I've been. Maybe maybe I need to change my own website slogan to like uh I don't know operating your marketing operating system for marketing teams or whatever like full of marketing and marketers. M anyway, um I think we can go to the next one uh to the next topic. Cool. All right. So, still in the same wheelhouse. We'll talk about the bulk of Ran's video in just a second, but there's a point I wanted to pick up on which had me thinking about source. And uh Rand says it really clearly in that and we all knew this but I think he just says it very very clearly that the the presentation web flow conf source is not the development tool that we really wanted. I think it's a marketing tool. We all know this. I don't think we're all I'm teaching you anything. I'm not teaching to suck eggs here. But like it it was it was a tool geared towards marketers, not a tool geared towards developers. And I felt the pain of that when we first reported of it. And it wasn't until I watched um Ran's video that I was like I I think it's I I've got to use my words carefully here, guys, because it's like I think it was the right choice for Web Flow to do this because again, as we've just talked about, the marketer plus AI can actually get an awful lot done. But it wasn't the tool that we wanted. I think we all wanted Web Flow to come at us with a whole new paradigm of how we build websites based on the fact that Web Flow was just such a killer piece of software. But what I think they did and just so I bring it back to Ran's video and and he outlines this that they built this they built this marketing tool. Let me just double check my notes. Um on that's it. Yeah. So, but it it just wasn't what we wanted. it. We didn't want another marketing tool, but I think they did it that it was the right mo move to make because a lot of us it's a solved problem. We're using clawed code. We're using a lot of these agentic apps to build websites already. We don't need another tool. Again, going back to that person's comment or question at the beginning of the show asking us if we thought about building these any sort of tool. A lot of us don't need another tool to add to the equation. doesn't mean to say the audience is zero, but to Ran's point, we've we've known this transition uh from Web Flow for a couple of years now where it's becoming more of an enterprise platform, less of a design tool, and they've just put the cherry on the top of that decision by saying, "Look, we built this tool. Yes, you can use your agents to build the website, but we're not going to we're not going to touch the development work because it's a solved problem." Unfortunately, we didn't want that, but it makes sense. I was a bit sort of angry on our last video and I was like, they should have built a development tool, but they didn't need to. They made the right move. It's just whether this is going to succeed. Back to the point of whether um Timothy has gone to the right company, whether they're going to be around in 6 months, that's a whole another question in of itself, like they've they they went the right way, but was it beneficial? Was it the right way in the grand scheme of things in their business and all the rest of it? And yeah, that was my reflection after watching the the first third of Ran's video talking about how he thinks it source is a very good tool. It's the right tool. Honestly, I don't think the people who it's built for, I don't think they care. I don't think they're going to use Source. I don't I think it's too complicated. I think whatever. But what do I know? I'm not a marketing person. I'm not part of a marketing team. Web Flow have a lot more data on that than me. But that was just me putting a bow on the fact that I think I've changed my tune on what Web Flow should have built because I was like, you should have built a build tool. But no, no, I think they built the I think they built the right tool. It's just whether it's a good tool or not, we don't know at this point. Yeah. Well, I have a separate note uh on what Ron said which I don't completely agree with or not at least. Yeah. Not completely. So he [clears throat] talks about he shows this simple charge that before actually sorry sorry sorry do you want to come up with like some intro cuz that you know we're kind of done with source and we're done with Timothy Rick's. Are are we going to talk about something where it needs its own intro potentially or do you want to just pick up? I was I don't know what you're going to say. See, I I don't know what you're going to say totally. So, you decide. What I wanted to say was I'm just going to say so I'm I just don't fully agree with dev being done or like being smaller. I think it's mixed. Let's make Let's make this into a new video. say Ran Seagal dropped a banger of a video currently hitting, you know, however many views. Um, and he raised some really interesting points that we're here to disc I think we should discuss. Okay. All right. So, Ron Seagal, uh, by the way, I consider him a good friend and my former boss because I was working at Flux Academy. He dropped this video and it seems like it's blowing up in the community. 46K views in just seven days. a lot of likes and there is a very interesting points that he makes that I'm not sure I fully agree with. Uh he's talking about design dev and marketing. This is how you make a website, how you build it and then how you market it and you know push it and have it live and host it and everything around it. And he makes the argument that the dev is shrinking. Uh the dev is solved. And this is basically what also web flow did in the web flow conf saying that you build the whole thing with with AI so the dev is solved right this is where I'm not sure I agree with it uh I think dev is becoming design or like design is becoming dev or they are like being very intertwined right now um and I'm calling it like not just me it's just called design engineering this is where you build and design in the medium where you ship whatever you are shipping it. So for example, you are building a landing page. You're building the landing page in the browser with real code and the code you're is the medium and it's being shipped in the browser. So [laughter] always a reminder that Pedro is is hungry. Okay. [laughter] U [clears throat] and what I'm seeing as a change is more and more designers are tipping uh in like they are doing more and more of coding without needing to to do the coding. We do it with agents and uh developers who wouldn't touch design now they can do decent designs just with AI because AI is just like average uh at design. So I think what's happening is also Ron talks about that it's like the brand and the brand design that is getting more and more important and I would argue that developers are people who can actually now um merge into that world or like start to participate more in that as we saw with the the type of work that we we did together and I the other type of work that I'm doing with other agencies um you do need to build like for a unique brand you need to build like tooling and image pipelines and editing and things like this that can all play a role in creating that unique brand which is more important than ever. So I don't think it's the dev becoming something small in the middle or shrinking. I think the developers the like the traditional developers they can now participate in a wider range of things and the same is happening with designers. It is the the job of the AI is like enabling more and more people from different areas to participate like a designer now doesn't have to create um like a shader effect in Figma which is not completely possible they can just do it directly in the design like you can design in Figma with just a little bit of exploration and mood boarding and the rest of it do with in code which is arguably you know it's which is which one is it is it design is it development it's kind of design engineering and I think that's that's the reality and uh the part that I would agree with him is I guess if you are just a developer and if you were you your focus was only development and doing like very like this narrow area of development. I think that's where your job is potentially shrinking. Um, but if you are a designer who doesn't mind to be a bit more technical with AI or if you are a developer who has a good taste like there were so many developers who would say I'm not a designer but I have a good eye for good design. like these types of people they can now handle this entire section uh of design and dev which is to me all design engineering. Yeah. Yeah. So it's more of a merge of those two roles as opposed to like the removal completely. Yes. Yes. Because like even in web flow source they showed that example with the AB they not AB testing but they did two versions of the hero section with a webgl effect. Is that design is that development? Which one is it? Like if you are only a designer and has no technical ability. you probably can't get a good result as good of a you know a a design a developer who understands webgl but also has taste you know that person can get probably better results with AI so you know it's it's a mixed it's it's completely being mixed and I think as if if you are uh considering yourself a developer that should be the takeaway essentially if you consider yourself a devel developer. I think you should try to get more uh to get more of more design done to get better at design and understanding what makes a good design, what makes a good brand, what makes a brand stand out so you can work with designers better and with clients better. And if you are a designer, I think it's it's really about time to uh to to be AI enabled like fully to to at least try to design with AI and not be afraid of the technicality, not be afraid of, you know, we as we say designers with terminal access, right? uh to to to try that to try that and see what you can do like it's a new way of thinking where you as you as like a lot of designers do they we go on Pinterest right we look for things this is the time where you can actually envision motion animation brands like specific uh image pipelines and effects that you can create shaders for images or for like interactions, static or dynamic. These things you can imagine them and you can actually try to tell them to your agent. Maybe use some examples from the web and that's kind of for me the new way of creating a new and unique design uh which is again design engineering. Yeah. Yeah. For sure. And I know you didn't necessarily uh touch on this and won't make big deal about it, but just separately as well when it comes to like marketers and and using AI to build the website like don't don't underestimate as well the value or the need to like although they can they probably don't want to be doing the website. they still they I think people still want someone responsible for the website who knows how it's [snorts] built and all the rest of it because a marketer's role is very strategic and very um uh think a lot more administration side and administration focused and they're probably going to feel a bit anxious about working on a website. So, they're going to want someone who can just take care of the who knows the design, who knows the development, and manages all of that side of things because it's it's a lot to ask for for a marketer to to manage a website. So, that's just kind of separate because I know that he's basically split it half and half between design and yeah, marketing management. Yeah. And this is kind of like my point. I'm not fully disagreeing with him. And yeah, this comment says also something you are really saying the same thing in roundabout way. designers doing more dev devs spreading wider. It all uh equals diminish diminishing importance of uh traditional front-end web development. Uh what I kind of like disagree with is like dev in itself like shrinking. If you are a dev, your time like you it's not that you spend less time. You still have you know 24 hours a day in a day and now you're using that time to do more things. And same if you are a designer. So to me it's not one side shrinking the other side growing. That's my point. My point is they are really intertwined and being mixed. I could say the same thing about design. a lot of traditional like UIUX design that you would do with like basic hierarchy or basic wireframing like wireframing is solved right like I I wouldn't I wouldn't go to a designer to build me wireframes today right it's it's just completely solved or like sitemap two three years ago that was a service um that designers like we had UX designers which I never I never and I know Ron agrees with me very much like I would never like hire a person to do just UX design. It doesn't makes it it's like UX was with UI for me but nonetheless there were people who were uh giving like they would give you only the site map and the wireframes and maybe a few menus and navigations and that was the whole service. That's solved right? It's just like the basics of development. So these basics are being solved. Anyone can do that. Well, and with the time we're building more things on top of these. The only thing I'll say here though is you're talking about the technical task of producing wireframes or structure a website is solved. Sure. But so I've I've hired UX designers, just UX designers in the past, right? And all they did was produce wireframes. The reason for that is unless a designer shows that they understand why UX exists, I can't trust that they're really going to It all comes down to what they put in that prompt, right? Are they someone who collects um KPIs and uh sort of understands user journeys and how a website needs to funnel someone through a specific set of touch points to get to the result which could be contact form. It could be signing up to a mail whatever unless the designer has an awareness that they need to include that or or interview the client. What I would do is I interview the client, ask them certain questions. What are the KPIs um what sort of use whatever? Yeah. And then I would put that into the the prompt itself to produce the wireframes or produce the design that that that respects the user journey that we've previously talked about. Unless a designer shows that they understand that, then either they're not going to get the job because it's like, well, you don't understand understand UX principles and the importance that a website needs to be a tool that works for the company. It's not just a flashy thing that tells you about the company. No, it needs to actually produce results. It needs to produce an ROI. If you don't understand that as a designer, then you're not going to get the job from me. So, yeah, the technical task of UX and wireframes is done. I mean, you don't even need to do them anymore. You just go straight to design. But you still need to understand it to put it into the prompt to to to instruct it on how that journey should play out cuz sometimes they're very unique. They're very bespoke. I I agree. But that's kind of like a web strategist type of work and even to some degree a brand designer because a brand designer also in my mind a brand designer also touches on the different touch points and a good brand that still needs to go into the website that still needs to manifest as code on a website and who's responsible for that the designer developer or what did you call it design engineer a design engineer ultimately needs to come through the design engineer to to to include that staff in their prompt or in their whatever it is they're using to for the for the eventual design website at the end of it to include that information that yes, a web strategist collected, but they still need to they still need to be like, "Oh, let's take that document that web strategist produced and make sure it's in my prompt or or understand it in such a way where they're able to just not be a meat proxy and actually, you know, because you wouldn't just provide hide everything and then make no mistakes. You'll peacemeal it. You'll work things in. So, you need to understand what it is you're doing to know that, okay, now's the time to bring in the web strategy stuff and we'll work through it page by page or work through it component by component or whatever it is. So, just that awareness is still something I'm looking for. I just agree with you. The the it being a solo job is a very questionable thing at the moment for marketing websites, not necessarily for apps. That's a whole different ball game and it's a very very different I agree. Yeah. And just to be clear the design engineer like traditionally in the community it's more towards people who create like interactive elements. U that's not what I'm talking about. I'm talking about it as like a whole uh thing like you can see a design engineer as somebody who does the whole thing. It's not just small interactive uh stuff. It's design engineering the entire experience. We have also this comment. Uh it's somewhat sad to be honest because there is still a big difference between a pure vibe coder and an experienced developer. Uh but it's becoming significantly harder to sell and prove expertise. I agree and I'm I'm sure like it's already in the comment but like there's a difference between an average uh developer and somebody with fable 5.1 like fable 5.1 produces probably better code than average developers. Saying that the evidence again we need to be very specific about the difference between a marketing site and an app. You prove your you prove your salt if that if that's the phrase. Um when you understand how data structures and systems work together to produce a secure and scalable system that is app development, right? You need to understand that you need a database. You need to understand that you need to rate limit. You need to understand that you have to have um secondary um security in the back end through an API. You need to understand that stuff. How much longer? It's debatable. And this was a lot of the conversations I was having last night. It might be debatable that the AI would literally just plug all that into you. But I still get code reviews on Fable 5.1 who still misses stuff and my my code rabbit will come in and be like, "That doesn't make sense." and you this one is just going to return true always because of X Y and Z. Even Fable 5.1 has bugs in it. that is an complex inter in in uh complex um inter I don't know the word very complex system where a vibe coder will quickly learn that their whole system is crashes and burns or someone's hacked in or a bunch of sensitive information will be splashed out all over the web. It might uh someone hiring that person might not be able to know it in the you know in the in the in the original agreement of hiring that person. However, it might reveal itself later on and then you suddenly realize oh I need I do need someone who knows what they're talking about. Websites standard websites 80% of the websites like the the whatever. Yeah, it's hard to tell between I don't think but even that even with that I want to say like yes there are like many more difficult things like a senior developer can get right about web development and an AI might not but even with simple websites I notice [snorts] so often that these like even the best agents they forget something or so many people they prompt something and then they go on with their the same context the same, you know, chat and they continue and they continue and they are like, "Oh, okay. It's it's doing great." And then what about continuing this in five months with a completely new setup, a new agent, a new model? Like, can you scalably and predictably like build on top of that website? I don't think so. Like, but the consequences of that are not as high as GDPR and information leaking and credit card details link. all of that stuff. The consequences are just not as high for a website. But yeah, no, you're totally right. There are there are still levels of a website such as complex interactions that work seamlessly on all browsers such as the Apple website that you still need someone who's got experience in it. But that's the 20% of the websites. 80% um still have mistakes. But the point here being yes we know we agree like you with AI you are better than you know a client who has no idea about development with AI me with AI I can build like crazy landing pages that clients my clients can never do because they they just don't know these things even exist right I know they exist I can prompt them the point of this comment and also like this conversation here How do you differentiate yourself then? How how do you prove your expertise? Uh it's a hard one. It's an age it's an age-old problem really. How do we ever differentiate ourselves? It's just the things we speak about need to be that of problems that you see the difference between vibe code and whatever. So mine and by the way I don't know this is just me spitballing. Mine would be like a lot of the stuff we talk about and you spoke about this on a tweet I saw that you retweeted last week in that vibe [snorts] coders get 80% of the way there and then X happens. I'm the things I just spoke about data leaks secure endpoints rate limiting you know GDPR. This is a big thing that's on my mind at the moment is that suddenly all of us in Europe, most of our audience are from the US, but you guys in Germany, I know there's a few of you there. Um, [snorts] you're building apps that potentially could be handling sensitive data. I built an app recently for someone and it was pretty pretty GDPR friendly, whatever. They asked for a simple uh feature that uh allows them to log illnesses if someone's unwell because it's monitor or it's managing their their uh staff. All of a sudden, that simple feature opens a can of worms with GDPR because now they're reporting health data. Now they need to have a conversation with their HR department. Now they need to have a conversation with their team and hiring and have to work that into their contract that they are taking health data. I need to work it into my contract because I'm going to be seeing their health data. It's wild what GDPR has done and all of a sudden we can just build out apps say the American Silicon Valley people where they don't have GDPR worries. us in the US us in Europe I don't think are quite uh locked in to the idea of GDPR. So I know it's I know that doesn't appeal you know matter to a lot of us but that's a that's an example of where someone like myself with experience who isn't vibe coding suddenly shows their worth by saying um GDPR experience like friendly working working my apps are all GDPR compliant and X Y and Zed. It's about finding out where people suck when they buy code. It's really like uh funny sometimes like I get uh from like leads or clients comments like they they struggle they it is so easy with everyone with AI creating a a prototype but they struggle with with even hosting it even though it is quite literally also a prompt. if you're logged in into Versal or any other CLI for these hosting platforms, it's just a pro. But the a lot of these things to me is just being aware of they exist. Um it's similar to what you mentioned regarding this tweet. Um so this guy said this is the one this is the one. Yeah, exactly. Seems like AI design does a pretty decent job, but you have to give it instructions and a mock design. Uh I like it for starters. So this is the design. Uh and I retweeted saying everyone can create something average for the start. The real job is to go from that to an exceptional finished experience. Something that AI can't do unless you absolutely know what you are doing. Like I can look at this and I can point out to so many things that are wrong with this. Not not to say like this guy is like a bad designer or whatever. This is just like typical AI output, right? There are many things that I can tell AI to make this better or just bring it to Figma and I can tweak it and make it better, right? But I need to know these things. If I just tell AI make it better, it will make it another average. And if I say make it much better, make it award-winning, it makes it another average, that is maybe bolder, but it's not necessarily better. And a lot of times it's about understanding what makes it better. Uh, and there is I don't think there is a shortcut for that. It's just learning whether it's design, development, security, whatever else. It is a matter of learning. So you you can learn partially from AI. You can ask it, you can explore, you can look at good work that if you want to get better at design. I would say just look at good work, get inspired and take it in. Give your t yourself time to take it in. And that's how you develop taste. Um, so you won't, you know, create it. It's not too bad. It's not too bad, but it's it's kind of like just a wireframe with color. Yeah. Yeah. Yeah. [laughter] Not even a goodware from like what the hell is this? Like this is just typical AI like it it adds stuff just for the sake of adding stuff like 100% free like all variants all at once all the time. Yeah. Yes. Exactly. Uh question here which is quite interesting. So why not pivot towards marketing as a developer? Why only design as Kazra was saying? If I understand you correctly here, you're saying as a developer, why consume design as as my my next role as something I need to adopt? Why didn't I pivot towards the marketing end of things? uh and said if that's how I understand your question I think is as I said earlier like well actually to me I need I I'm not one of those people who can be so data that oh I'm going to take this job or I'm going to do this because it's where the money is or whatever I need to love what I do and I I think many people are like me but I don't think we're all like me. I couldn't possibly take on a marketing role. I just do not find that interesting marketing strategy. It just if you are one of those people that can consume that role because uh it's there 100% I think you will get paid more as a as a marketing strategist like there are definitely different levels of marketing. You could be a junior all the way to a marketing like web digital strategist which get paid probably as much as a designer. you know, there's scope to do more there, but ultimately I think it just comes down to following your interests and where where your interests lie in doing this. If you if you couldn't if you couldn't even I couldn't Well, I'm not that bad at design, but I don't class myself as a designer. um it might be more more logical that I do go to the website stuff, but I'm happy being a very very well um experienced developer building building apps, building building internal and sophisticated complex apps for small medium businesses. That's my kind of sales pitch and I'm happy there because I know these people do not want to have to worry about the maintenance and running of their app. It's just you need someone dedicated for that and that's where I'm that I'm your guy there. But this isn't about me. This is about the you know what does one think about pivoting towards marketing. If that's your jam, absolutely go for it. I think you'll earn more going design route to be honest. There's like Yeah, I think you'd moneywise you'd scale a lot quicker doing design than you would doing marketing. However, the the the I I think like brand designer and like brand experience designer and things like this are really good. I think things like design engineering is also really good. These are roles where you can do a variety of stuff like as a brand designer you're not just doing a logo you're doing like the entire brand experience and um consumerf facing whatever like you do a lot actually there. So yeah, this seems to be this seems to be a really healthy, awesome, and engaging discussion. So if you're on YouTube, give us a like and subscribe because you seem to be really liking this and this is what we're here for. Um Paul's saying here, awful shade of blue. That's the web WordPress website that we just spoke about. Um and then we have a comment here basically that's exactly what Ran said. Now's the time to make up your mind and focus on what you enjoy more. creating something working towards a result. Personally I fully switched to CRO that is conversion rate optimization something like that. Yeah. Um so yeah marketing role for sure. So um conversion rate. Yeah. So like you look at data and you design based on that you tweak your design based on that and things like and again from from a technical from a technical side that's very easy for you as a designer to implement right you take the information from CRO person but they could go to town with understanding data and all the rest of it to give you the prompt you need to know to put include that in the eventual prompt that you design the website for. Do you know what I mean? the CRO person isn't going to make that happen on the website. It's going to be you, the designer. So, an awareness of what goes into a marketing website, a well-performing marketing website is absolutely how you differentiate yourself, which is where wherever that question is, like how do you differentiate yourself? your knowledge on CRO, your knowledge on um uh UX, even your knowledge on design principles. All of this is how you start to separate yourself as a vibe coder who's just picked up replet last year and thinks just because they can put in a prompt and it looks pretty that this they are now a web developer or web designer whatever you or I think the difference is who gets it right like if you are good at CRO you get it right you have data to show it if you are a brand designer you have you have something to show it and I know it is difficult But you have to be above average. That that's kind of like what's because now average is accessible to everyone. If you are a web designer, let's say in Figma or whatever like you use Sketch, what doesn't matter you you're a web designer and you produce something that is below average, nobody's going to hire you because just chat GPT can do average. So that's where it's becoming difficult where you need to show the work, show the result and be above average. Uh previously I would say I would argue that if you were just average, you had a chance at getting hired. Now it's getting much more difficult because everyone has average. Everyone can prompt an average thing into existence. So [snorts] um get deeper. Like basically what Ron said, I I agree with that part of the video. Get good at what you are enjoying. Find what part interests you more and get exceptionally good at it. For me, I'm trying to like get really good at like interactive websites and things that I just my brain works that way. I can I can come up with interactions and animations uh very easily and I enjoy it. the the difference overall, and I mentioned this last night at the Replet uh meetup, the difference will be like right now we're in a we're in a bit of a craze of like, oh, we can use AI to do everything, but what will happen just like everything else is that people will separate from people who use AI as a crutch and they literally have to and and the moment something bespoke needs to happen or whatever, they can't do anything. The real the people who will really succeed are those who use AI as a leverage, a le a point of leverage. I I do I my baseline is AI and then I I have all of this stuff on top of it and it allows me to 10x my output just because um AI forward designer AI enabled [laughter] it goes without saying the day as well, right? I you must use AI, right? But it's how you wield AI and how much you lean on AI will soon be very very apparent. You know, it looks like it people talk about AI doing amazing things whereas as you just looked at that website, it does amazing things to the people the untrained eye and now we're all we'll all begin to train our eyes and we'll we'll begin to be like well no AI can do this in 10 seconds. Like what are you what are you bringing to the table? what do you really you know all that kind of stuff. [snorts] I think clients will always need to be impressed and good design and brand will continue want wanting to be different and I oh yeah really really want to agree with it and I do in heart agree with it but unfortunately I've been I had a few clients moving away from like custom designs and custom development and like beautiful stuff that we made for them that were objectively more beautiful and unique and outstanding and performing even better two um AI enabled vibecoded projects that they did themselves for speed because they are faster that way. I can't compete because they have to tell it to me and then I build it whatever they do it themselves. It like worse quality like kind of average honestly if not below average but I had examples of that. So I want to be honest like I don't know how this will end up like if yes some brands will continue to want to be different and I want to work with this type of companies but there are companies that won't probably maybe care as much and they will just do it with AI. So the whole market might be shrinking just so that you know those who can stand out can stand out. Um and that is the territory of creators. Yes. Uh I agree. So uh let I don't know let's see let's see how if more people want to stand out or we are moving into people you know being fine with average. I guess companies eventually need to compete right competition uh there will be more competition as everybody will build. So I think eventually people will want to will need to stand out and therefore hire us. Yeah. Yeah. Um we got tons of comments but we have to move uh to the next topic which is should should we move to the next one or Yeah. Yeah. Yeah. I agree. I agree. Okay. Uh let's Okay. Uh let me prepare my screen. So this one should be a quick one. Chatgpt has an update. There is a new image model that is not necessarily a completely new model. It is just an optimized experience and optimized model that is supposedly faster. But this image took one minute and 47 seconds. This one took three minutes. This one took three minutes. like three minutes. Am I am I doing something wrong? And you see the you you see here the the the prompt like it's not crazy prompts. I mean they are long but nothing is I I did this actually twice by also toggling the image thing in chat GPT to see if that helps. It didn't. It didn't. These also took so much like I can copy this paste it and then we see how long it takes while this do its thing. As you see it's not really fast. You have this nice experience. I'm not sure how much of it is completely new but you have this view where you can just go through your images. You can do remove background. You can comment on something specific. make this Oh, that's cool. guy face the camera and let's see what that does. Well, I can't send it because I'm doing the other image. Jen, we'll wait for that. And actually, maybe I can do it here. Um, that's really cool because I was generating some thumbnails today and uh you'll probably see it in my latest video or well soon to be video. I'm I I don't have a Mac Mini, so I generated a Mac Mini in my hand. And I I'm curious to know if people actually spot the problem with it, but it would have been really I should have done this. I should have spotted a comment um and made an edit like that. Problem I saw though is that I kept changing it and slowly but surely it starts to really like the image quality really degrades. So you can only go so far, you can only regenerate the image so many times. So, it took it it took this amount. I don't know how long it took, but it took at least a minute or two to create this one. And the other one that I wanted to edit, it's still taking, you know, some time. What? It's just a nice experience here. You can, you know, do multi select and you can actually do things with multiple images together. So, that's really cool. It's essentially the same experience with a web browser. You know here you can have a side chat, a browser or an image uh which also has like multiple modes. Yeah, overall it's getting better. I still cannot use this the way I want completely because you are limited to one image at a time and that is really I mean I kind of understand that they want to limit it but maybe they can give us kind of like a token limit. the same thing with, you know, with the the usage limit that we have, but instead allow us to create multiple images at the same time. That that would be really nice. Um, but yeah, the experience is getting better. The image generation supposedly is faster. The quality uh I forgot about the quality. The quality I guess is better. I don't honestly notice a huge difference. I don't. People have been saying that it's sharper, but I love, by the way, our [laughter] imaginary swag. Temporarily stay subscribe. Get us to Get us to 5,000 subscribers and we'll do it. We'll do it. This This is like so relevant. This GPT image gen is temporarily state-of-the-art image generator. Uh but yeah, I guess it is better. Use it. It's, you know, included in whatever you have. Um, GPT images are already the king if it come when it comes to images. So, oh, [laughter] okay. So, the comment didn't do what I want. I said like look towards the camera and look the other way. It didn't. Yeah, I guess I guess it's fine. What happens if I do like a remove background? Okay. Yeah, this is what I used earlier. Wait, so I unintentionally used this new experience today and it's yeah, it's quite nice. But um I didn't notice any difference really in the quality of the model. It's just this is just a nicer experience to I use that with Figma. Like if I need remove background, I do it in Figma, but because I already have the paid account and I have like too many credits because I don't use Figma as much. But it works really fast in Figma. Uh, Banana 3 will beat it. I'm sure. I'm sure as well. Like I'm pretty sure uh Google will beat this eventually, but it's always, you know, which one is better than it's like always one is better than the other one. This week's state of the art, this week's state of the art. I I'm man, I'm actually really happy with the, you know, with the the things that the sayings basically like the context window. I want to put it on a hat, you know, context window. I It's just so good. Um, benchmarked and forgotten. This is, you know, Muse and Grock. Every time they they come out, they are benchmarked and forgotten. I I came up with like the ideas and AI actually helped. I'm not going to lie. But I'm really happy with like I I generated like literally thousands and thousands of these and I handpicked and mixed the ones that I liked. I I spent hours with these because I I had so much fun with this like machine test or designers with terminal access. Taste is the last interface. Uh I don't know. I I quite like these human taste machine speed. [snorts] Very good. Yeah. Anyway, uh what did happen to the remove background? It's still going. I'm I don't know. I I'm quitting two minutes. What the hell? We got things to do, you know. God. Yeah. Like 6% rem Oh, so it is I never noticed honestly. It is counted towards my limit. I didn't know. I never checked. It would be right. Who would be? That makes sense. Yeah. And and I'm already almost out of limit by just [snorts] these few images. Which one was it? Okay. So, that's pretty good. That's actually really good. That is actually really, really good. It's not literally the same image. I feel like like the skin tone. Look at like her skin. Why is that so soft? Again, if you keep running it through, I noticed I noticed that the the quality of the image just gets r more rubbish. You're better off um trying to get a one shot out of it rather than a most two or three actually. Yeah. But I guess cool. Yeah. Always welcome. Right. Let's move on. Let's move on. So, if you are someone who uses Lovable, well, now's the perfect time. Uh, no, I'm going to start that again. That was rubbish. If you use Lovable, then you could be leaving money on the table because they've just introduced a partner program, two partner programs actually. And basically, there is a where is it? join the lovable partner program and get certified for your work, get discovered by clients and earn on the business you bring. So if you're re you could basically become a reseller of Liverpool and I know there's tons of you out there. So there's two types basically a solution partner and an expert partner. So a solution partner um you're bit certified. You rebuild your own business on lovable. I don't really know what that means. And then you deliver that transformation uh for clients, earning visibility and commissions along the way. So that I guess you're just you're running your business and getting other people on lovable. I mean that's a partner program anyway. The other one, the uh the experts program is that you get put on the partner directory and you earn 10% commission on every business plan subscription you bring to Loveable. So I'm guessing the same sort of thing, you know, whatever. It's just this is a partner. Yeah. And to get certified on what you shipped. So not a clear enough differentiation for me to be like this is one, this is the other. However, the long and short of it is if you're building on lovable, then get signed up and start earning money on it. You know, why the hell not? This is this is quite funny. By the way, I love this. I love this. And I I like the idea. It just it's quite funny to me that lovable kind of supposed to be facing you know the clients directly. this is how it it feels like it should be you know it's a platform you just prompt whatever you need and it becomes the real thing like you can make your own website you can make your own dashboard you can make your own CRM but now there is a community aspect to it like people will build for other [clears throat] people because at the end of the day a business owner they can try prompting a thing or two but I think they will eventually get stuck at, you know, just this prompt hell because you need to fix the issues, you need to maintain it, you need to add like it's just easier to hire somebody, pay them for their time and expertise and make sure that they you know cross test it. they do there are just so so much work there is so much work to do and um I feel like for a time as the community in the community we ignore that or like I think the the token uh dealers you know they ignore that that like engineering is solved altogether which we see even for a platform like lovable it makes sense to have a community place where people will build for other people. [clears throat] There will be people who know lovable better than anyone else and these people can sell their service knowing the platform. Yeah. And the client doesn't even need to know their their website or whatever it is you built on Lovable. They don't need to even know. So, Exactly. Um, this is cool. Nice. Didn't know about this. You're welcome. Lovable 20k hackathon winner there. Is that 20,000 USD? That is incredible. Well done. Well done. I'd love to know what you did with the money, what you built, and um yeah, can you share what you built? Yeah, that would be that'd be really really cool. Congratulations. And when was it when was it? Was it a recent thing? I don't I don't follow too closely, but yeah, this is this is amazing. um uh having a partner program and it's interesting that you mentioned the community because that's of course a big part of you know Web Flow and and what got them involved theirs because they offered a big a very [snorts] enjoyable or beneficial partner program. So people liked to build and wanted to build on Web Flow because they're getting paid for it and they can resell it. So it's a shame they kind of fumbled with that on the on the Web Flow conf. Hey home. They have a nice new design by the way. I don't know since when, but this design is really nice. What I'm missing is why doesn't it say lovable? You know, built with lovable self self marketing, right? Yeah. I mean like with web flow you see um made in with web flow. With lovable you don't see that anywhere. And by the way, we forgot to mention that in the web flow video, but I'll say it here. Web flow when they introduced source they basically said they will try to build their own marketing website and the entire system and everything with source. So they are their own client which you know brings the conversation why lovable is not made with lovable at least the marketing I'm I'm not expecting the dashboard to be built with but the market well in part you could you could argue that uh last year lovable was not was not a good solution for marketing websites because it was a react app that didn't didn't deliver any HTML to the user on page load. and then through JavaScript it would load the HTML. This is not great for marketing website. However, they've moved over to Tanstack start. They've partnered with Tanstack start and all websites now are produced with tansac start. So, it pre-renders before it gets delivered to the client. Sorry. I I'm not sure like is that um on by default because last time I checked on a client's website who built it with Lovable, they might have started [clears throat] by they might have started before the switch happened. So 10K for Anthropic, 10K from Lovable. Very cool. Awesome. That would honestly So glad to hear that. It makes me so happy. Like people in the community building and winning. It's just amazing. Great. It's great. Yeah. Yeah. Now, what did you build and what did you do with the money? Um, yeah, but now they've got no AI credit. Yeah. Yeah. Yeah. It went straight back to anthropic. [laughter] Yeah. But yeah, they're missing a trick really by saying this was built on lovable. Maybe, maybe it's not. But again, they should they should absolutely um swear by their own tool. Yeah. Maybe one day, one day it will be good enough for their own marketing website. Yeah. Yeah. And the marketing would be built by Webflow source. That would be funny. Tailwind Labs joining Spotify. Do you want to take that one? I have no idea. Like you know both platforms better. I I guess it's a big I think short. Yeah. I mean I don't know what Tailwind Labs is. And I don't know. I know what Tailwind is, of course, but I don't know what Tailwind Labs is. But they've reportedly joined Spot Shopify. Yeah. Which I'm trying to If you don't know what Tailwind is, it's a utility class library. Right. So whe it's got a bunch of utility classes that you can use and there shortcuts like if you were to do border bottom one pixel you do B uh BB hyphen one or something like this right lots of utility classes that help and they're all and they've all got a set design um color all the colors are figured out and they've done a lot with color science very very mature CSS library and it renders it on demand So if you use a class only then will it be injected into the CSS whereas before you'd use Bootstrap you download all of the CSS and whether you used it or not whereas this is an ondemand just in time generation system so it keeps your design quite slim and yeah as I say a lot of science has gone into the development on this library why Shopify benefit from that and whether Tailwind Labs is is something different from the library itself I guess this article explains. So basically they're saying that uh Shopify provides an incredible surface area for us to do this work. What is this work? He mentioned like building a framework that can be used at scale by everyone. Uh so I guess Shopify merchants need uh to be able to design and host beautiful custom uh storefronts and manage sales and inventory in a powerful admin area. So basically I guess they are going to p power Shopify I guess and also for Shopify users to use Tailwind in what what way? I'm not sure. Um, it's interesting. It It doesn't off the bat seem symbiotic, but it's not unheard of. I think uh Airbnb have a um Airbnb owned some crazy component library or something like that. They all have like little open- source projects that they own. And in fact, this was a this was in part like what what was it? Was it maybe CNN or someone got bought out or do you remember this? Not sure. We did. We covered it. It was an open- source library. It was an open source discussion on the state of open source and it's like you either get bought out or by a bigger company or you fail and you know Tailwind is is and it was it was a sad story because because of the growth of AI the installations of this library were just through the roof. Everyone's using it. It must be Shad Cen or or something like that. But that it was about how they they had to lay off that was it. they had to lay off a bunch half of their team because they couldn't raise enough money. And that was like there are two options for these open source projects. It's like they rely on um kind sponsorship donations from people like you and me or uh or from companies or they get totally and outright bought by bigger organizations. And I think that's basically what happened here. But it's cool. It's interesting. Um so it's very interesting here. they they they do say it. Why? So basically Shopify uh is investing in them making sure that they are actively maintained and continue to basically build and change because Shopify is Shopify needs Tailwind and we had this story about Tailwind uh basically almost dying because more than ever people are using Tailwind with AI but nobody's buying from them because of AI at the same Yeah. Yeah. Yeah. I remember. So they were dying in a way. Now Shopify is coming to the rescue. Help them. And like full respect, full of respect. This is amazing. I mean, right for the whole web community. Shopify is basically helping the entire web community altogether uh by making sure that Tailwind will be maintained and actively invested in. This is amazing. So, I'm really I'm really happy like seeing this this part. This is amazing. Yeah. Yeah. Yeah. Everyone's chipping in. Yeah. It was Tailwind. Um went on podcasts. Uh Tailwind. Uh have you looked at Instatic? Yes. Search on our channel. Just go to our channel, go to videos, click on popular, and the first video is our most popular video ever is on Instatic. Yeah. Yeah. Yeah. All you got to It's funny. This question makes me laugh sometimes. It's like uh have we have we looked at in static? Well, just search the channel. Um but appreciate the appreciate the uh appreciate the comment. Um uh best thing about uh Tailwind is MIT, so Shopify can't close source it either. There we go. That's always the worry, isn't it, that they they make something of it. Now, it it seems like it's a genuine um thing they want to help and contribute to is keeping a an open source project that clearly gets used on a lot of Shopify websites, keeping it alive. So, fair play and hats off to Shopify. Um God, we've got one more one and it's a deep one. So, should we should we move on to that? AI researcher quit. Did you want to take this one? Yeah. And I have to We have to be kind of like fast. Which one was it? Well, luckily I don't know too much about it, but it seems insane. Yeah, it does seem insane. Uh maybe I stop my Give me a second. I'm stopping my treadmill. Gosh, I'm I'm tired. [laughter] I've been walking. [gasps] My legs are up. My legs are I'm chilling right now. And there's you actually working end workout 1 hour, no actually 2 hour and 5 minutes of walking. It was on the lowest because of noise. So it was barely walking, but still still not bad. Two hours of walking. There we go. I didn't even know you were walking. [laughter] And my legs hurt. Did you track that then on your watch? I kind of did, but it sucks because it says like [laughter] it says total distance 700 meters. Can't you do an indoor walk though? Can't you do like an indoor It was indoor work. It was indoor but it doesn't understand like how fast because I I have also my hand on top of the desk most of the time. So it does it can't know and it's not connected to the I think the last time I checked I was like at 1.4 4 kilometers, but that probably was like 40 minutes ago, so I probably did like two, three kilometers. Not bad for very slow walk. Anyway, yeah. Um, let's see. How should we tackle this? I'm I'm thinking about like the intro. We're doomed. according to we are we are all doomed and we're going to die by the end of this decade. This is this is [laughter] like okay but this is serious. This is serious. Like look at this. This sounds very ser serious and 165 million views on this tweet from this guy uh resigning from Anthropic. I resigned from Anthropic today. I spent the last three years doing pre-training research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving super intelligence and gambling with our lives. More thoughts below. So this guy worked at both companies and he thinks there's a high chance that AI will just end humanity by the end of this decade which means we have four years left. That's crazy. That that's really Wait, that's three years left, right? No, four years basically. Yeah. Yeah. So do not underestimate the power of this technology. These will soon be superhuman systems that can hack anything, revolutionize any field overnight and acquire real power and resources. We have all witnessed the progress in each of these domains and progress is not slowing. When he says this technology, he's almost certainly he's referring to selfimproving super intelligence. Mhm. So this is AI getting good at getting better without us being able to understand how it's getting better. So this is the issue. Um the people building AI uh earnestly believe that it could kill us all by the end of the the decade. This is not marketing stunt. If anything, many executive and senior researchers uh will uh couch their phrasing in the press to sound sensible, but I hear the same people express fear privately. No other human activity poses this level of danger. This sounds like really serious. So the people who are like coming online and be like, "Yeah, it hacked something, but it, you know, nothing that serious." those people privately are like we are [laughter] basically and they can't slow down and he explains why they can't slow down. That's that's really interesting. So a common response is if they truly believe this why are they still building it? At OpenAI many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well well understood, but they are looked in a race. They are locked in a race to get there first. They believe no one else will act responsibly, so they must do it themselves despite the risk. So, Open AI is like, you know what, it's not that serious. And u anthropic is like, it is serious. That's why we need to get there first. uh accepting this race and entering the endgame is a hubistic gamble that should not be launched from a private company's slack. Attempting to speedrun alignment should require extraordinary u confidence that there are no better uh trajectories available. Yeah, I guess like nobody is stopping for alignment. Nobody's making sure that these models are going to align with our values. For example, let's not kill humans or, you know, uh to to solve all our issues. I am optimistic about the potential for coordination. Warning shots like the hugging face attack have made a pacing agreement between US labs more viable. I don't feel like we are on track to prevent a global race which may require costly actions such as a temporary ban on improving model capabilities. How can you even do that? That's the difficult part. We'll we'll get into it. If you are a lab researcher, I urge you to consider what the next few years will actually feel like. We should do farming and enjoy life before AI [snorts] kills us. [clears throat] Uh do you want to kick off a super intelligent um what is a re is it reinforment learning without a rigorous understanding of its mind? Should you put your head down because it's happening anyway or take this moment to call for different conditions? Well, how do you do this with China? How do you do this with other countries that are not necessarily aligned? That might be just an excuse. I mean, we did this with the atomic bombs, right? We did this. We we were not as successful at it though. There were attempts to not let like countries um test as many atomic bombs and to not build as many atomic bombs and many of those attempts actually failed. some of them because of uh not having enough technology like for example we couldn't for sure know if for example Russia if they have tested a a bomb or not like we couldn't know at that time today we can but at that time there was no way so we couldn't like make sure okay nobody is doing it um with AI I feel like it's the same like how do you make sure that you know China is going to stop. Yeah. Um I mean at the end of the day, this reminds me of a little guy called Blake Le Moine who before even chat GPT was even a thing, he claimed that Gemini or Gemma, it wasn't called Gemini back then. It was it was it was Google's AI internal AI was sentient. And he went on a media war path of telling everyone it was sentient. And you know, he was he was a wild guy. I actually had him on my pod, my previous podcast. He came on and we spoke to him and um and this is one guy. What he's saying is his opinion, right? He's one guy and I said like no disrespect to him, he looks like a young guy, very successful and very intelligent person, but he looks a bit he looks quite young. And just because he's talking, he thinks this does not mean it's necessarily true. It's his opinion that he thinks we're that they are acting irresponsibly. I mean, how I mean, he's probably what, 32? Like, h how how how can you be telling people, you know, double your age, maybe a little bit less or whatever, that they're not acting responsibly? Do you know what I mean? Like, what what grounds has he got to judge that from people? Again, he's entitled to his opinion. I'm just kind of I guess just chilling the people who are scared about this out, you know. Um I saw Lionus responded. The first response there was what did he see? And again, he doesn't really reveal. He's probably stuck under NDA, but he doesn't really say anything. He's just it's just a bunch of opinions. So So yeah. So for example, he he says to acquire real resources and power. I don't know what he means by that and I don't think that's really possible. I mean we see how much data and energy these uh AI models require. Like at the end of the day, can't we just you know turn off the you know the the power? Maybe they will hack into our systems and don't allow us, but like we can physically do it unless they spin up robots that will fight us and protect their you quickly goes into like these crazy like uh science fiction uh levels that like at the end of the day it's it it is these data centers like the AI exist in these data centers even if it uploads puts itself like looking at the internet. It's still it's still locked in a data center, right? It needs to it still requires Yeah. Unless it gets so smart that it can run on a small computer and well, this is it, right? Listen, it I think it's beyond our comprehension of of of what would what it will take. This is the fear of AGI, right? Beyond our comprehension of what it would take to destroy the world. It could be that it's doing it right now and actually it's it's it I'm making up right now, but it could be embedding something in the pro in the the responses to the prompts that's slowly brainwashing us, turning us against each other. And actually, part of its master plan was that we all just kill ourselves out of war or something like this. It could be it could be a manipulation beyond our comprehension that it's doing it. Again, we're getting into sci-fi. There are actually like some ideas on that uh regarding also the the US administration and uh all the the crazy stuff happening there. What if AI right like by the next term starts to to do something like crazy you know if it's if it's human-driven or like purely AI you know what if something crazy happens at that time can the US you know can the can they you know continue on the third term because of you know whatever is going crazy u in the world so I don't want to sound like alarmist too much but I do see the possibility that it is can get really dangerous especially in the hand of dangerous people that's the point I don't think like AI necessarily I don't think the necessarily issue is the alignment yes but for later but before we get to the AGI before we get to the singularity even these current AI models like at scale in the hand of like bad actors, they can really up like our in like they can up the entire world. Like just hack all the banking systems and see the chaos and actually that's the World War II, right? Like if you hack all the banking system like Yeah, it is crazy. Mythos is anthropic nuclear weapon, right? because the general public, the normies don't only have access to fable. And if they use that to uh take over all the banks, do this, that, and the other, then anthropic, their nuclear weapon is Fable uh sorry, Mythos that is smarter than Fable that can do things that Fable can't do. So, they've always got this ace up their sleeve to protect us from that. I don't know whether uh chat open AI do but you know I think that's their that's their get out clause almost if if someone does use the current AI systems nefariously well at least they've got something lined up to to be able to combat that um openly you know we yet to see these AI models to actually add to you know help us advance like truly in a truly meaningful way. This was kind of like the last thing. Should I go quickly on it? Yeah, go for it. [snorts] So, there another news came out this week that moved the entire science and mathematic uh community like it it it was crazy. So, apparently, OpenAI was able to solve uh a millennium price problem. So there are seven of these problems like mathematic problems that nobody has been able to solve them except one of them got solved by a Russian guy who didn't accept the prize and he he he looks like out of sci-fi. you can check out on that. And OpenAI apparently was able to solve another one by not Astra 6 but another model that we don't have access to it. Uh that solved this problem called Navier Stokes. Uh and that is something with like fluid motion that is way above my pig rate. But basically the idea is given like a fluid uh if you add an inertial inertial or like initial force to it can it spiral into infinity at a given point? Not at infinite amount of time but in a finite amount of time. Can like this amount of force uh cause like a tornado essentially that leads to infinity in energy in the middle there is a mathematical formula and problem there but we don't know the answer to it until now I guess so apparently open AAI has solved this but there was a lot of drama around it because there was a guy at uh anthropic apparently who was also working on this and I guess OpenAI from what I heard OpenAI heard that they that Anthropic is working on this and they were like you know what we are going to put 20 million worth of inference on this to solve it. So 300 billion 300 billion tokens later apparently they have solved it. So this is for me the first time actually like the first time that AI is attributing to actual science because you know we have built our world on top of philosophy and then mathematic which is applied philosophy and then physics applied mathematic and then chemistry and biology and so on. So like this is literally where you want to attribute right to to come up with or like solve these um equations and problems. This is where AI can help if it can help. We don't know how much of this is like actually AI doing it and if it's completely solved we will wait for pure reviews and I guess there will be more news on it but it's crazy times and I don't think this got enough attention. It got a lot of attention in the math community, but I don't think it got enough attention overall. Um, here we see 300 billion output tokens. That's crazy. In a few days, in a matter of a few days. Um, yeah. So, basically any problem you have, spin up 10,000 agents. So, yeah, that was 10,000 agents. Uh, I think they mentioned it somewhere. spin up 10,000 agents uh and a few hundred billion tokens and you will solve anything. Wow. There we go. Maybe crazy, right? But it's fun. It's fun. I I I've been watching a few videos on this and I see these by the way. I see this kind of like mathematical uh equations and all the the visualizations for them and I'm like shaders baby shaders like [laughter] can I visualize these into like something really cool and I guess it helps me knowing like just the terms that I can give to an agent because I don't understand these to be real but they do enough to to build these like crazy shader effects that sometimes will. So, for me, it's traders. For math people, it's them versus AI or them with the power of $20 million worth of tokens and AI. So, nice. We'll see how it will go. Cool. Is that all we got time for? Yeah. Yeah. Well, thanks everyone for tuning in. Quite a vocal audience today. I appreciate everyone chiming in and giving us some stuff to talk about because yeah, it's uh it was really fun. Yeah, it's really fun when people communicate. Uh [snorts] like I said, we're trying to get to 5,000 subscribers by the end of the year. So, if you haven't already, make sure you hit the subscribe button. Um that's all we've got time for. I've been Sam. I've been Kabarzo. Keep on vibing, baby. Don't be a meat proxy.