were re-implementing a lot of the same capabilities in various different contexts. So my assistant that I had in my coding editor had to get a web search tool associated with it, but so did claude. ai and so did all these other services where we want Claude to be able to interact with the outside world...implement a set of functionalities once and take that everywhere. So build it once and configure everywhere. So the same web search functionality that I might get on Claude Code will be the same web search functionality that I would get on Claude. ai. - Okay, that makes sense. Creating that universal compatibility basically when we have tons of applications that
Поиск по транскрипциям
Visual Studio Code integration, and it works really well. On top of that, you also have the Claude agent SDK, so you can actually build custom AI agents inside Claude Code and then you've got different ways you can use it. So you can use it inside
back to Claude Forson. This is interesting. So, B new an awesome tool for code generation, right? And so, they've said that Claude Sonet is our go-to model for code generation workflows and it's using the 1 million context window, right? So, you can actually build with this. If you don't even want
they do officially launch, they'll have thousands of hours of real world testing already done. The competitive landscape based on benchmarks Grock 4 beats Claude Opus 4 on reasoning matches it on coding accuracy destroys everything on speed costs more than Claude but delivers more value. What's coming next? August 2025 XAI launching a dedicated coding model. September
benchmarks again it's comparable to some of the top models they include deep seek R1 and grock 3 which recently rolled out here too and they also introduced Claude code now this one is only available in research preview right now so you may not have it yet but because Claude Sonet 3. 5 sonnet...coding and frontend web development capabilities so we'll test this a little bit here too but you can see the default writing is so much better if you've ever used chat GPT or even Gemini they go overly promotional still to this day they do improve that all the time but I still prefer Claude with his default
going to zero. Cursor all them. They're all going to zero. Lovable. Uh the the business model makes no sense. The only winners will be Codeex and Claude Code. The business models simply do not make sense for vibe coding rappers. Why is that your opinion? Because the vibe coding rappers, which there's hundreds of them
writing code faster. It's about automating entire development workflows that changes everything. Okay, quick recap. Anti-gravity is an agent first IDE. It plans and orchestrates your projects. Claude Code is the execution engine. It writes and tests the actual code. Together, they automate full development workflows. You give them highlevel instructions. They handle the rest. It's fast
look because you can select your repository, connect it to your GitHub, and it will appear on the left-hand side in the Claude application, kind of like this. So there will be like a Code section, but it has not been released to the general public. So it's kind of like how on the ChatGPT mobile...update. That is the only time I really find myself using it on the go, but I'm sure many other people have different use cases as well. And Claude Code will be releasing a web version of the application soon as well
know, we'll end here, but how do you know, I'm just curious about Obsidian and Claude Code in general. Uh you don't need to show your Obsidian, but because I know it's a very personal vault, but how — I can show it. I don't care. — Okay. How do you think about using as a creator...using this kind of folden folder nom and like folder structure in nomenclature. So I would just immediately ask like cursor or cloud code. Cursor is a bit faster if I use cursor with like claude. Um but yeah, I really think it's kind of like a superpower for anyone who's writing their stuff in files
also removes stale context as well. So, I think that's particularly useful when it comes to coding because there may be a particular terminal output that happened in Claude Code that you no longer need, and that can be removed, and then you don't have to compact the conversation quite as much. There is some documentation about
video coming out in two hours. It's the best models to be using with every cloud code use case and the cheapest models to be using with every claude code use case. So you can decide, okay, I want to use the most expensive best models or I can use the most cost-effective models
actually out on a walk when I recorded this. So, you know, sort of just like brain dumping my uh my content production process and uh and getting Claude's feedback on that and then trying to kind of hammer out what this would look like to have a team helping me in the research and planning and production...this is where I can actually work on them and then continue to refine them using cloud code. So here is a conversation that I've been having with claude code to rework and improve some of the processes that are built into the skills that my agents use in our task system. And so once I have, you know
test out some of what I've been doing here with uh just my um vibe coding tool, creator scorecard. Yeah, I built this with a combination of Claude Code and Manis, which is probably now my favorite AI tool. It's probably Manis now. Um, yeah, we got a lot of great AI tools in the ecosystem...think that's enough to take it to full production grade. Um, I would think a minimum of a 100 hours into uh vibe coding this might be sufficient to take it to production grade. I'd want to probably do 100 to 200 hours on it. So, I'd want to give it another 3 to 6 months. Then
script MCP server. You can set up with Python. You can set it with Node. How do you set it up? You just paste this code into your MCP settings. Just go to Claude desktop. Go to Claude. Go to fire claude settings developer and then edit the config and insert the MCP code inside there. Right. Where
like a plugin. What do you even describe it as? — Remotion is basically a service that makes video editing or generating any components or any motion graphics into code. And because Claude, especially the Opus 4. 5 model, is so good at generating code, it can take advantage of this new Remotion service and create any videos you want just...that? — Yeah. So ultimately if you think about it um what are the video editors doing right and video editors in the back end is just code. So what you can do with reotion is when you set up a cloud project let's say you dump in all your let's say your a-roll your other uh stat
person had just like vibe-coded it really? - Yeah, that's correct. So, they developed ransomware, kind of like worked with Claude until eventually they got Claude to actually write the pieces of code that they wanted. There was a lot of refusals that happened and Claude said no, and they would provide some sort of justification for what they...stand up a carding service. Meaning the actual infrastructure you'd be using to conduct some sort of scam, you'd be using carding infrastructure that is run by Claude. Again, I think it goes into the theme that every stage of a scam or a cyber attack or an act of fraud, and there's many stages. There
actually coordinate sub agents in the same way that you do that on open core. You can do that inside claw code as well. So the key difference is really between open claw and claude it's like claude makes this easier. It's a bit more friendly for nontechnical people. There's no technical setup whereas something like open...need server configuration docker API keys etc. So that's the difference between both of these and that's why I'm beginning to use uh claw code and claude in general a lot more than open. So thanks so much
couple weeks back where Anthropic dropped their newest models, the Claude Opus 4 and the Claude Sonnet 4. Watching the live demos, especially how they used it with Claude code, their command line tool, was pretty crazy. They showed extended reasoning, tool use, and real-time code edits. Opus 4 is the world's best coding model with sustained performance
Today, I'm going to show you the most powerful free AI coding tool that just dropped. Uh, it beats GPT4 in coding tests. It beats Claude in real world apps, and it costs nothing to use. I tested it with five crazy projects, a snake game, a portfolio website, an expense tracker, a story generator, and a productivity timer...making Open AI and Anthropic nervous. In its release announcement on social media platforms X and GitHub, Moonshot claimed Kimmy K2 surpassed Claude Opus 4 on two benchmarks and had better overall performance than OpenAI's coding focused GPT 4. 1 model. But wait, there's a catch. The RAW model is slow to use directly. That's why most
homepage for itself. Claude opens Chrome, searches for claude. ai, and then types in a prompt asking the other Claude to create a personal homepage for itself. Claude. ai returns some code, and that gets nicely rendered in an Artifact on the right-hand side. That looks great