Replit's NEW AI Agent 3 Just Took The Lead! (Here's Why)

Vibe coding platform Replet just took the lead. It just took the lead. So they announced agent 3. This is basically an overhaul of the platform. The AI itself. Replet is an is a vibe coding platform and I really like it. I'll leave links down below. Um affiliate links just to FYI. I really like it. I've often started a lot of my apps inside a replet because I feel like it's really great for just architecting uh architecting an app like getting it getting it up and running without having to worry about all the individual components. It's really really good for that. So I built the AI I really [ __ ] hate AI app inside of Replet and you're seeing uh agent 3 right now actually. It has everything built in. So you've got databases built in which other platforms don't. And it's a Neon database, so you know, still third party, but it integrates really well with Neon. It's got authorization built in. Um, it's got secrets, which is a lot which is what a lot of other platforms like Lovable, like Bolt, all of these other platforms, they don't really have a lot of the integrations. I describe this as the web flow of Vibe coding coding platforms. It's got everything built in. You don't need to worry about anything. I've since move I mean this is still in replet but other apps where I take it beyond the prototype I then export it bring it into clawed code and do this that and the other I always criticize replet because it it was very overwhelming look at all these tabs that I've got going on it just you know what I mean it's very and if I just if I actually just keep closing these maybe this this is what I was going for it starts to get a bit starts to get a bit more familiar with people from lovable where you've got the you've got the main agent space here where you'll just chat chat chat chat chatting and it will just build it out. But you also do just have this tablete thing which is unique to replet where I can start to dig into not only some of the features it has like assistance secrets authorization storage publishing all that stuff but I can also look at files and actually go into the code and start it's it's just a bit more like nuance you can really dig deep into it but the simplification that you can see right now is is something new to replet 3 it's got actually app testing built into it so and with this max autonomy feature which you basically just trust it. It's like yolo mode in in cursor or whatever. You just can let it run for up to 200 minutes. Not quite sure how they're doing that from an API standpoint which I think they use uh claude. Yeah. And it's a really like just a welcome addition. I think they've tackled that aspect where it becomes overwhelming. It's still from a design perspective, I still don't think it's as friendly. And I'm talking about UI rather than UX. Um, still not as friendly as some of like lovable and stuff like that, but I'm telling you, it's it's by far and a way a lot better if you can just get over that aspect. It can do a lot more for you, I think, as a as a vibe coding platform. So, and what what was also quite interesting and more from a broad perspective in this live stream is they're trying they're trying to encourage you to think more and more about building these apps to encourage your day-to-day usage of AI. So rather than again going to a corporation, buying a subscription to a SAS app, build your own frigin app, you know, and they they were reiterating Matt Palmer would just reiterating this of like just building these apps just to just to help you through your day and encouraging that me mindset and mentality, which is what Claude did with archives or and no artifacts. Artifacts. Yeah, artifacts. you know what they're talking about just building applications to help you get your work done for yourself. So that was quite an interesting conversation that they were encouraging through the live stream. But overall really really good uh really good um set of features. Uh and it comes off of the back of a $3 billion valuation raising 250 million in a series D round of funding. So they're they're on the up. You know, I think Riplet is a really cool product for the next generation of makers, people who are going around finding problems and, you know, trying to solve them. And I believe it's not meant to, you know, to be used to solve problems for millions of people or even hundreds of thousands of people. Couldn't agree more. It might be actually solving a problem for three people. And I actually mean it for three people. So imagine like there is a company of 20 or company of 50 and three people on a higher you know position they they use a few different products to do something specific about their business uh and it bothers them and we have enough example of this like the just go to any business and see what software they use it's absolute trash. So what uh next generations of makers I I like to call could do is going to these businesses, finding these problems and building a prototype or even a working MVP for them and even might be able to finish it off and just package it and give it to them working just fine because it doesn't have to be scalable. It doesn't have to be to have like the the cleanest code. it doesn't have to have all the standards if it's something internal used by you know a few people to solve a very specific problem. So I think for that riplet is amazing and it can actually lead um to you to become like a engineer because that's where you start and then you understand like you learn problem solving and you just take it from there you learn from the AI and I can very well imagine that you start with replet and end up with something like cursor within a year. I mean like I keep encouraging you to use you know you're you're you're very you're stubborn in using cursor and wanting to use cursor but I think these are great platforms to learn the ropes of architecture if you're if you're willing to really you know I like you can max you've got max mode show you max mode yeah but with cursor I have just so much control and I have this like ease of mixing cursor with web flow and I I like to keep you know the host posting on web flow for the majority of the things that I do because I'm already paying I already am familiar with the platform and I have the max control over the looks you know if I have if I want to change something in the looks uh the looks of what I'm building I just change it directly in web flow and the functionality is what I'm building with cursor that that's why I'm so stubborn with that the only the only thing I I would you know tweak about what you were saying you were talking about makers going into companies solving the problem for three people. I their encouragement was really looking at yourself and looking at your own problems to solve because you all have different needs. You also all have different requirements and if yes it takes a certain mindset to to you know build yourself a tool. I I want to build a tool for my cocktails that I make uh on the weekend. So for my for my family and friends well like an app that they can order. I want to have like my own order. Like I I I have a specific way of uh making it and I want to to do it my own way. Well, so everyone can chill out and just order a cocktail and it would come up on with their phone and like uh Yeah. And I will make it and then I ask my brother to to deliver it to them. So half the time, you know, we've got everyone's got all these problems. You I what I was going to say is you need to be tuned into realizing, oh, I can build an app to solve these problems. And we're certainly in tune with that world. But their their their message out really was like anyone just look at your look at what you're doing. Look at try and think of how an app something you do repetitively the automation aspect just how can you how can you make your life easier by building an app. You asked me a question last week and a lot of my focus on the responding to that was like what's their need? Have you spoken to them? Whatever. I think our industry is completely um saturated with people trying to read the minds of society where truth is they haven't got a the foggiest idea. So you need to go out and you need to speak to people and find their problems. The the the chances you're you know oh I've got a great idea for an app and you think it's going to be good and blah blah blah blah but you've not gone out and spoken to anyone about it. The likelihood that app is going to succeed is next to nothing. You know, we're we're in this delusional world at the moment, which again going back to this video by Dave Shapiro sort of alludes to is like we're so insular right now. We just we think the world is kind of like um well just not going out and speak to people, just not going out and trying to solve problems for real people. We think that we can just come up with it. And low key, I don't want to go into it too much, but I've been thinking a lot today about the quality of developers that have gone completely freelance versus quality of developers that have actually worked in an organization. And I don't know about you in design, but for me, developers who have worked inside of an organization are leaps and bounds like better than those who have just always done um freelance because they've just never learned anything outside of their own comfort zone or little bubble sort of thing. Um, so again, don't know where I was going with that, but it's just this idea of um I don't know where I was going with that, but there uh just nattering on as per usual. This is part of a larger conversation on my show, Command AI, which we stream live every single week. We discuss the news and all things related to AI in the world of design and web. Catch us next week and join in the banter.