Being creative requires taking risks
Way Enough
Gas Town’s Agent Patterns, Design Bottlenecks, and Vibecoding at Scale
This feels like one of the most critical, emerging footguns of liberally hands-off agentic development. You can move so fast you never stop to think. It is so easy to prompt, you don’t fully consider what you’re building at each step of the process. It is only once you are hip-deep in poor architectural decisions, inscrutable bugs, and a fuzzy memory of what you set out to do, do you realise you have burned a billion tokens in exchange for a pile of hot trash.
Your App Subscription Is Now My Weekend Project · Roberto Selbach
John Henry in the AI Era
Learning in Public, Building in Parallel
Why Cowork can’t work
Running Claude Code dangerously (safely)
The Code-Only Agent
We're All Software Developers Now
Agent Psychosis: Are We Going Insane?
Don't waste your back pressure ·
Anthropic made a big mistake
Why Designers Can No Longer Trust the Design Process
Don't fall into the anti-AI hype - antirez
2026 is the Year of Self-hosting
AI is a business model stress test
Notion AI: Unpatched Data Exfiltration
How to write
Shipping at Inference-Speed | Peter Steinberger
The Unbearable Joy of Sitting Alone in A Café
Creativity faucet: Increase your creativity
It’s hard to justify Tahoe icons
Computation's Limits: What a Neuron Teaches Us
January 2026 | Maggie Appleton
Too Fast to Think: The Hidden Fatigue of AI Vibe Coding
You can make up HTML tags: (Maurycy's blog)
Why I Write (And You Should Too!)
Codex vs. Claude Code (Today)
I know you didn't write this · ammil industries
A non-technical attempt to explain 2025 LLM-based ai