Keep reading
More Articles
Why Models Are Getting Dumber on Purpose?
Reasoning scores keep climbing while per-token compute keeps falling. GLM-5.2 scores 99.2% on AIME 2026 with about 40 billion parameters active per token. Qwen3.5 hits 91.3% with 17 billion active. DeepSeek V4-Flash runs 13 billion active. For scale,…
I Analyzed 700 New Software Projects by Indie Hackers. Here's What Everyone Is Actually Building in 2026
I pulled the numbers from Forg across 700 public projects, 1.5k+ builders, and a mountain of company data to answer three questions I keep arguing about with other founders: What are indie hackers actually shipping? What stack are they running? How a…
Why I kept building even though nobody cares
If you're a builder, you probably know the feeling. You pour your heart into an idea. You spend days or even months designing, building, debugging, and polishing every little detail. Throughout the process, you're convinced this could be the project…
The shared system under every device in Copy: Walkie Talkie
Nine devices, two builders, and a system that needs to work when someone outside Copy designs the next one. Originally published on Minimum Viable Thoughts , with the full set of images and video clips. Copy is an async walkie-talkie app my friend Ma…
