STOP making tools for other nerds! GET THE FUCK OUT OF THE COMFORT ZONE! Make an AI tool that would help your MOM or DAD or someone who doesn't spend the whole fucking day in front of the computer.

Would you accept an offer from a company that pays you more than your current package but doesn’t allow you to use any AI tools?

For my own stuff, I’m pretty much done with CMSs. The last few sites I’ve built are just plain PHP, markdown files with YAML frontmatter, and Parsedown for markdown-to-HTML rendering. No database, no admin panel, no plugin ecosystem accumulating technical debt. Takes just 2 seconds to deploy and nothing to update, nothing to break, and nothing to maintain. Every WordPress site I’ve ever run has eventually had a plugin conflict that broke a bunch of stuff. Every Craft CMS or Drupal upgrade has had at least one “why the fuck did that break” moment. And for what? So I can edit content through a UI instead of just opening a file and pushing to git? All that just so bots can hit my admin login page? Ugh… I get it for enterprise clients with marketing teams, structured content, permissions, workflows, etc… a CMS earns its complexity at that scale. But for a personal site, a blog, a product landing page, or even for relatively large sites, markdown files pushed to git is the most liberating stack I’ve used in a long time. I’m switching most of my content sites to a simpler system like this. My co-founder is considering moving his main site over from WP to a system like this. Like thousands of pages, variations, modules, components. All that stuff can be replicated pretty easily, and then never have to deal with WordPress plugin updates ever again No opinions to fight, no framework to learn, no dependencies to resent. Just files.

Building an homelab is adultness Building an homelab is adultness Building an homelab is adultness Building an homelab is adultness Building an homelab is adultness

Chat interfaces require way more cognitive work than people realize. I find it tiring sometimes. I mean, you have to remember what you want, translate it into words, guess what phrasing the AI understands, do the whole back and forth dance, and track the conversation, summarize it in your head, then maybe branch out, etc. Traditional navigation just needs you to recognize a menu item. We need both, but man sometimes a chat ui is just not necessary.

listen to podcasts instead of music.

Every tool I love using has one thing in common. Someone made a decision and stuck with it. No settings page with 40 toggles and no “customize your workflow.” Just “this is how it works, it might not work for you and that's ok.” I don’t want options, I just want the thing to do what it says on the tin. I wish more founders had the guts to build like that.

Today seemed like a good day to clean up my phone.

🚨 BREAKING: Someone just rebuilt the entire AI assistant stack in Zig. It's called NullClaw. The binary is 678 KB. It uses ~1 MB of RAM. It boots in under 2 milliseconds. No runtime. No VM. No framework. No garbage collector. Just raw Zig. Here's why this is absurd: → OpenClaw needs a $599 Mac Mini and 1 GB RAM → NanoBot needs 100 MB RAM and Python → PicoClaw needs 10 MB RAM and Go NullClaw runs on a $5 board with 1 MB of RAM. Same functionality. 0.1% of the resources. Here's what's packed into that 678 KB: → 22 AI providers (OpenAI, Anthropic, Ollama, DeepSeek, Groq, etc.) → 13 chat channels (Telegram, Discord, Slack, WhatsApp, iMessage, IRC) → 18 built-in tools → Hybrid vector keyword memory search → Multi-layer sandboxing (Landlock, Firejail, Docker) → Hardware peripheral support (Arduino, Raspberry Pi, STM32) → MCP, subagents, streaming, voice, the full stack Here's the wildest part: Every subsystem is a vtable interface. Swap any provider, channel, tool, memory backend, or runtime with a config change. Zero code changes. It even encrypts your API keys with ChaCha20-Poly1305 by default. 2,738 tests. ~45,000 lines of Zig. Zero dependencies beyond libc. 100% Open Source. MIT License.

we will know we have achieved agi once you can one shot a solidworks clone that matches its functionality nearly 100% with a single prompt that runs overnight.

i built an entire x86 CPU emulator in CSS (no javascript) you can write programs in C, compile them to x86 machine code with GCC, and run them inside CSS

🚀 Gemini 3.1 Pro vibe-coded a double wishbone suspension • Independent double wishbone design • Dynamic coilover shock absorber • Vented disc brakes with performance caliper • Real-time kinematic travel & steering simulation AI isn’t just generating visuals anymore!

Asking Gemini 3.1 to make an educational app about Blob Counting Algorithm. One of the foundation algorithms of computer vision. The result is really good, one prompt, easy to test directly on the aistudio and the simulation of the dark/light behaviour is super realistic.

Everyone wants to "build an audience" but nobody wants to say anything that might piss someone off. That's why most content sounds the same. Safe takes get safe engagement. Strong opinions find your people.

I'm falling behind. I'm lost. I'm confused. It's becoming clearer to me that agentic AI is the next big thing. All your work done in the background. All Internet interactions done through a prompt. All tasks managed by specialized agents. I already have trust issues with LLMs writing my code. But now I have to learn to trust an always-running agent with access to everything. And it feels like if I don't, those that do will surpass me.

Introducing GLM-5: From Vibe Coding to Agentic Engineering GLM-5 is built for complex systems engineering and long-horizon agentic tasks. Compared to GLM-4.5, it scales from 355B params (32B active) to 744B (40B active), with pre-training data growing from 23T to 28.5T tokens. Try it now: chat.z.ai Weights: huggingface.co/zai-org/GLM-5 Tech Blog: z.ai/blog/glm-5 OpenRouter (Previously Pony Alpha): openrouter.ai/z-ai/glm-5 Rolling out from Coding Plan Max users: z.ai/subscribe