About 15 days in and already burned the max x5 plan 😓

It’s a bold move, we just haven’t solved the code bloat generated by AI yet, and it’s bloating exponentially fast...

Steve Jobs: “My model of business is The Beatles” “[The Beatles] were four very talented guys who kept each other’s negative tendencies in check. They balanced each other, and the total was greater than the sum of the parts. And that’s how I see business. Great things in

Just building stuff 😅

Local claude-code fallback 😅 qwen3-coder-480b-a35b-instruct-mlx with gemini-cli: 270.10 GB RAM and 262144 content length

86k tok/sec with qwen3-coder-30b locally, now that's something !

A steep improvement for local models usage with the chatgpt open-source 120b model, openai/gpt-oss-120b - context: 131072 Really the same kind of thinking than chatgpt o3, a bit more chatty and eager, but you know what to expect and this is a good quality.

Built a custom Claude Code status line that tracks my API billing in real-time: 25 sessions since billing reset, 55% through current 5h session, daily/monthly token usage + costs. Now I can see exactly where my money goes! DM or reply for code or tips :)

The GPT5(s) thinking efforts

How frustrating 🤯😓, impossible to have access to gpt-5-high on the official OpenAI API webpage… And no « gpt-5 » is not the real deal, the input limits are an order of magnitude smaller making it impractical for coding with CLI

Finally, time to test the real thing 😁

Thanks for the ride Claude-Code — 2.5B tokens this month! 🤯 Doubling down on Gemini for reliability. Adding GPT-5-High for brains. Pushing my tools, my efficiency ⚡️, and my limits 😅.

🚨 Protect yourself from supply chain attacks targeting AI coding assistants Malicious attacks like the recent "npm packages post script exploit" can now exploit AI tools in a clever way to steal your secrets: • SSH keys & passwords • API tokens & credentials • Crypto

Dang — I had just cancelled my Claude Max 20× plan after weeks of dual-testing it against GPT-5-Codex-High… and now they drop Sonnet 4.5 🤦🏻‍♂️

Finally someone says it 😌

Finally Gemini 3 🎶

The pace is accelerating! Claude 4.5 just landed, setting a new world record on the highly regarded SWE Bench Verified benchmark, surpassing Gemini 3 and GPT 5.1 while dropping costs to Sonnet 4.5 levels.

Agent Skills is now an open standard It's been great to see the traction Skills are already getting in the industry and this makes it easier for everyone to build and contribute to them🚀 agentskills.io/home

That’s a good insight; there isn’t much else to tell the model before starting a coding session. The rest depends on developer input.

This morning, I checked my phone, there was 150 000 AI agents connected to moltbook (reddit for AI), this evening they are 1 500 000 ! how crazy is that?