assistant that lives within any AI chat application that you already use. In other words, it's like glue between the apps and something like Chat GBT, Claude, Cursor, or VS Code, you name it. and it lets you control over 500 apps with just natural language. So instead of building an automation and struggling with connecting various APIs...mark promos as spam. And Rube just does it. It even remembers you across various applications. You just authorize once and then run Gmail actions straight from cursor, claude, manos, whatever. For creators and builders, this can be huge. I can pull comments and likes from various platforms, drop them into a Google sheet, and flag the top posts
Поиск по транскрипциям
choose your own adventure game. Start in a dark forest. Give players choices. Create different paths and endings. HTML only. Again, this tests storytelling plus coding skills. Once again, Claude 4 delivered great interface, smooth gameplay, multiple story paths that actually made sense. You could spend real time playing this and enjoying it. The user experience was incredible
Miniax 2. 1 is from China's Miniax and both are free to use as I've already shown you today and both claim to actually beat Claude and GPT on coding. Both have 200,000 context windows. But which one is actually better? We'll be testing them live today and running them head-to-head
something crazy to show you today. There's a new AI model that just dropped called Kimmy K2. 5. It's open- source, is beating Claude Opus 4. 5 in coding tests, and it's completely free. This thing can read images, write code, and think through problems like a human. I'm going to show you exactly
inside VS Code. I'm talking about the new AI toolkit 2. 0 and this thing is a game changer. You can now run GPT5 claude or any model you want right inside your code editor. But here's the crazy part. You can build AI agents that literally write code for you. I'm going to show you exactly...toolkit? is Microsoft's new VS Code extension that turns your editor into an AI powerhouse. Think of it like having Chat GPT, Claude, and every other AI model built right into VS Code. But it's not just for chat. This thing can browse models, test them, fine-tune them, and even deploy them to the cloud. Here
anyone watching at home I want to be % clear with you here right number one you can get free access to Claude number two it's got a higher context token number three you can preview code inside here which makes it awesome now if you want to use the extended thinking mode which already seems to be faster than...that mode but on the other test is scoring 84. 8% for Claude 3. 57 Claude 3. 7 versus Gro 3 beta so Claude is winning in that respect agentic coding you can't see any benchmarks for extended thinking but you can see no extended thinking and it's storming right it's 20 or 30% up right there
Claude loop removes all of that. You set it up once you tell it what done looks like. Then you walk away. Claude keeps working it. It checks its own work. It fixes mistakes. It improves the code. It doesn't stop until the completion signal shows up. Let me break down how this actually works. First, you write...task. Next, you set a max number of tries. This stops the loop from running forever. Usually, people set this between 10 and 20 tries. Then, Claude starts working. It builds the code. It writes the features. It does whatever you asked. Here's where it gets cool. A stop hook checks the output. It looks for that completion signal
what I've tested so far. I'm sure it'll get better, but it's just it's nowhere near the same level as Claude when it comes to this stuff. So, coding out this game, Gemini 2. 5 Pro one, when it came to coding out the first game, which was the 3D Runner game, Claude...subdomain. Claude can't and probably will never be able to do that. So, Gemini wins in many ways as well. Multimedia stuff, then you know, Gemini has it all inside the ultra package. Obviously, you have to pay a lot of money for that, but it is a powerful pack. So, we can see these coding out here
what I've tested so far. I'm sure it'll get better, but it's just it's nowhere near the same level as Claude when it comes to this stuff. So, coding out this game, Gemini 2. 5 Pro one, when it came to coding out the first game, which was the 3D runner game, Claude...subdomain. Claude can't and probably will never be able to do that. So, Gemini wins in many ways as well. Multimedia stuff, then you know, Gemini has it all inside the ultra package. Obviously, you have to pay a lot of money for that, but it is a powerful pack. So, we can see these coding out here
just jump into Visual Studio Code and I'll say you can now create sub agents. I'd like you to create a specific claude sub agent that reviews code after you're done writing it. The main agent will call the sub agent at the end of every major write process and use it to confirm that the code...just like always constantly be writing higher quality code. I have a feeling Boris has to do a fair number of like database calls and analytics based things in his day-to-day because he included an additional one called data and analytics. The takeaway here was just have Claude stop writing manual database queries. He's like I haven
Perfect. So that's coding with computer use and Claude. This took a few prompts now, but we can imagine in the future that Claude will be able to do tasks like this
code and that can tweak the project that you've built and improve it. And the reason that I like that is number one, Claude Opus 4. 5 is much better for coding and claw code is really powerful. But number two, you get limited credits with Gemini inside anti-gravity. So if you're using this program for free...going to exfically can anti-gravity control the terminal that's called code and control. I've never tried that actually. That'll be interesting to try and see what happens. Yeah, maybe it can. Another thing that you can do inside Claude, if you have a regular task that you do, you can actually set up something called Claude
code and that can tweak the project that you've built and improve it. And the reason that I like that is number one, Claude Opus 4. 5 is much better for coding and claw code is really powerful. But number two, you get limited credits with Gemini inside anti-gravity. So if you're using this program for free...just open up a terminal and then ask Claude to improve the outputs, right? I tend to find that gets you much better results. So for example here, if we say, okay, open up project in a browser, right? And this is something that we've actually coded out previously. This is a to-do list app. All right
abandon chat GPT or Claude. Those are great. But be model agnostic. Use the best tool for each job. Ernie for multimodal content. GLM for coding, chat GPT for conversation, Claude for long form writing. mix and match based on needs that maximizes productivity and minimizes costs. So that's the breakdown. BU Ernie 500 and GLM 4. 7 Flash
make our projects even better. So if we go inside the terminal here, we can use opus 4. 6. So we just type in claude inside the terminal. Make sure that you've got claw code installed and then from here you're off to the races and you can start using it right directly on your project...everything for you and you just tell it where you want to go, right? So I've built hundreds of automation systems. I can't code, right? I just describe what I want. Claude builds it. I use it and that's basically the game right now. So the solarreneurs winning with AI are definitely not the most technical ones
Claude to start up a server so that we can actually view the file within our browser. Claude opens up the VS Code terminal and tries to start a server. But it hits an error, we don't actually have Python installed on our machine. But that's all right, because Claude realizes this by looking at the terminal output
There's a brand new update from Claude Co today as they've released Claude Co-work. And this was literally just announced a few hours ago. Claude introducing co-work claw code for the rest of your work. Co-work lets you complete non-technical tasks much like how developers use claw code. So you can see for example...scrolling through, this prompted us to build co-work a simpler way for anyone, not just developers, to work with claw code in the very same way. Co-work is available today as a research preview for Claude Max subscribers. Claude Max, by the way, is like the top plan for Claude and it's the most expensive
tests general knowledge, this new Quen model scored 83. 0. Claude 4 scored 86. 6. That's close. But here's where it gets crazy. On coding tests, Quen scored 51. 8. Claude 4 scored 44. 6. Quen wins. On math problems, Quen scored 70. 3. Claude 4 scored 33. 9. That's not even close. Quen destroyed Claude
That's good for the entire AI ecosystem. The coding AI wars are heating up. We've got GitHub C-Pilot. We've got Chat GPT's code interpreter. We've got Claude's artifacts feature. And now Deep Seek V4 is coming in hot with these insane leaked capabilities. The next few months are going to be wild...biggest AI releases of 2026. And if the leaks are even half true, it's going to change how a lot of people think about AI coding assistance. So, here
mode, tell it to implement everything, and it executes. Files updated, functions created, tests written. This workflow is so much better than letting AI randomly change your code. Multimodel switching is killer. Claude giving you responses that are too cautious. Switch to GBT4 on the fly. Same session, same context, different model. Or use a local model for privacy. Open