Dawid Tomczyk

@antirez 🥹

There is another idea that continuously return in my head recently: we should not stop writing code by hand. But this code will be like poetry. Small programs that are outstanding either because of style or new ideas introduced. Code you may print and frame on your wall.

@Shreyas_Pandeyy @grok @elonmusk @cursor_ai Right now, the usage limits are low, and I suspect it costs more than Opus because of that. I've only been testing it for about 3 hours, but I've already burned through 30% of the weekly limit.

Man, I was so wrong about @grok . Sorry @elonmusk. I was proved wrong when SpaceX acquited @cursor_ai. This latest grok 4.5 is amazing and at this speed they can really compete with OpenAI and Anthropic very soon. Grok is affordable too. Great work @elonmusk and @cursor_ai team.

@techdevnotes Very low. 2h of usage and it's 20% already.

Is Grok Build’s weekly usage limit low or generous

Okay. Grok 4.5 is impressive and very fast. @grok

Built an XTermJS multiplexer running in an Android WebView, with a UI styled after ❤️Zellij❤️. No need to drag my laptop around anymore! I can create sessions on any PC (just requires an agent). All of this developed with @claudeai 😎. Claude has access to my entire infrastructure. Outside the home network it just requires connecting via WireGuard to a VPS with a public IP and OpenWRT at home

@unclebobmartin Are you aware that clean code no longer matters? 🤣 I read all your books (grateful for them) and spent so much time perfecting my coding skills... 😅 Can we laugh a little?

Very soon now, AI will infiltrate everything in our lives. We will tell our microwave to cook, we will tell our TV to pick a movie, we will tell our car where to go and where to stop, we will tell our house to get warmer or cooler. AI will screen our calls. AI will make our amazon orders for us. AI will organize parties and meetups and even just meeting a friend. AI will do our taxes, pay our bills, manage our finances. Etc, etc. Musk may be right about his notion of sustainable abundance. The 20th century lifted 1/4 of the world out of poverty. The 21st could be 10 times better.

Grok 4.5 in Europe when? Just bought subscription 😅 #grok @elonmusk

@ThePrimeagen Why not just focus on writing a good test suite?

Unfortunately no matter how judicious I am, no matter how much I review the code, it feels impossible to not let the slop slip in. I feel like you can prevent it from overflowing, but without your hands getting dirty, you don't really know the state of the project.

@mehmetaderin @Jilles I made another network 10.x.x.x on the second rounter downstream to my home router.

@Jilles How do you fully isolate it in your home network so in case of somehow somebody getting access to box the box is contained?

Please buy a small Minisforum, Beelink or GMKTech. Connect it to your router. Install Proxmox or even just Ubuntu and host some of your apps there. You can have Codex/Claude Code set it up and host small apps with Cloudflare Tunnels and Access. Here is what I run:

@claudeai Like it. Give me now the ability to choose effort before describing tasks, the ability to name agents and maybe group them?

New in Claude Code: agent view. One list of all your sessions, available today as a research preview.

@mweinbach @edzitron Most certainly it was on purpose hoping no one would find out. There's a reason they blocked other harnesses, they don't want to make the model worse so API is not affected but they can control subscriptions through Claude Code or Desktop.

Well this is sorta nuanced They didn't make the model itself worse, they changed things in Claude Code that made it worse Claude Code as a product got worse but that's not because they changed/quantized Claude. Until they found the bugs, they probably thought nothing was different. This isn't to say they weren't wrong, but this wasn't on purpose

it's wild that they did not acknowledge this the entire time - Boris even went around saying nothing was wrong! Laughable company, be very curious to see if this fixes stability issues

@MelkeyDev What a downgrade from 4.6. Very disappointed!!!

My current issue with Opus 4.7 is that the comment it adds are insanely verbose and useless. Worse than previous models

@elonmusk in 15 years money doesn't matter. Why are you building xMoney then? 🤔

A year ago, looking at Cursor and GitHub Copilot, I was impressed, but skeptical that AI was truly revolutionizing software development. If it were, where was the 10x more software we should see as a result? Then in December, @claudeai Code tapped me on the shoulder. 😊

The viral claim draws from popularized models of how neuroplasticity supports learning/habit change (inspired by research like Merzenich's work, not one single "study"). The 4 stages typically described: 1. **Awareness**: Brain spots new patterns or the need to change. Early discomfort hits here—most quit before seeing progress. 2. **Repetition**: Consistent practice strengthens new neural connections. 3. **Integration**: New pathways blend into daily habits/behaviors. 4. **Consolidation**: Changes become automatic and stable. Stick with it past the initial frustration; real rewiring takes weeks/months of effort.

Anthropic just mass-licensed 10,000 of the most influential developers in the world for $12 per person per month. The math: 10,000 maintainers × $200/month × 6 months = $12M in sticker price. Actual compute cost to serve these accounts runs closer to $30-50/month each, meaning the real spend lands around $3-5M total. Those 10,000 people maintain an open source ecosystem valued at $8.8 trillion in demand-side impact, according to Harvard and the Linux Foundation. 60% of them are currently unpaid. This tells you everything about how Anthropic views developer distribution. They’re acquiring the people who decide what tools get baked into every CLAUDE.md file, every CI/CD pipeline, every GitHub Action, and every project README across the most-used repositories on the planet. And they’re doing it for the cost of a Series A marketing budget. Think about what happens when a maintainer of a 5,000+ star repo starts using Claude Code daily. They write CLAUDE.md files. They add Claude Code GitHub Actions. They reference Claude in contributor docs. They build workflows that assume Claude as infrastructure. Every contributor to that project encounters Claude as the default. 10,000 maintainers each influence, conservatively, 50-100 downstream developers through their projects. That’s 500K to 1M developers seeing Claude Code embedded in their daily workflow within six months. GitHub spent years and hundreds of millions building Copilot awareness through traditional developer marketing. Anthropic is spending $3-5M in compute to get Claude Code embedded at the infrastructure layer of open source itself. The timing is surgical. Microsoft killed Azure Sponsored Subscriptions for open source maintainers in September 2025. Burnout rates among maintainers hit 44%. Quit rates hover at 60%. Anthropic walks in with the most expensive AI subscription on the market, handed out free, right as everyone else retreats. The selection criteria reveal the strategy. 5,000+ GitHub stars or 1M+ monthly NPM downloads. These are the people whose tooling decisions cascade through dependency trees touching every Fortune 500 codebase. $3-5M in compute for 500K-1M developers organically adopting your tool through the open source dependency graph. That’s $3-10 per developer acquired, embedded at the infrastructure layer where switching costs compound monthly.

If there's no human to read the code, there's no need for it to be readable. And readability was never the goal. Nor was maintainability. A correct end result was and is the goal.

@royvanrijn @gdb @paulg There is nothing wrong with that. I also took care of good design throughout my career. When I started playing with LLMs I even micromanaged their output code to match the style I prefer, until I recently realized it's not a scalable approach and soon you won't be able to compete

@davidtomczyk @gdb @paulg Taste in product design, definitely... But I really want the code to be clean as well, well designed, simple, etc....

I’m glad to be a Java developer with the rise of AI. With LLMs now writing most lines of code, reading will become *even* more important. Java, as a language, was designed to be readable and understandable, easy to reason about (static typing ftw). 🙌🏼

@royvanrijn @gdb @paulg Taste in product design, not code design. Designing clean code doesn't matter anymore. If you want to read code, you are the bottleneck. You just need to verify the correctness of the application utilizing a proper test suite.

@gdb Or @paulg:

@codewithsushi Forget coding. Focus on architecture.

Should I stop doing DSA (Leetcode and shit) because looking at the fast growing advancements in AI, coding will no longer be a relevant skill till I graduate.