Blog
2026-09-12
The Leaderboard Changes Every Week, the Excitement Never Outlives Criticisms
Excitement arriving in an afternoon lasts only as long as a frontier model holds its ranking — one or two weeks on average, and you never test the model yourself.
2026-09-06
"Ship Before Perfection" is out of date
The advice made sense when a working product took a team months. Now agents write the code, one person builds in a day, and what shipping fast mostly produces is slop.
2026-09-02
Fine-Tuning Is a Trade, Not an Upgrade
Everyone advertises what fine-tuning brings and no one talks about what it costs — mine bought key and instrument identity, and paid for it in timing.
2026-09-01
Apps Need Their Own Models, Not Prompt Shells
APIs are still efficient and cheap enough for everyday use, but for vertical apps it is increasingly essential to fine-tune a small model to serve a narrow area.
2026-08-31
LLMs Are Dead. Small Models Are the Future.
Today's LLMs win on text prediction and data volume, not depth — what we actually need are small models built deep in one field, with LLMs there just to wire them together.
2026-07-25
What Is Soulor AI
Soulor AI is a companion and social simulation platform — not just a chat buddy, but a space to build relationships and rehearse them.
2026-07-25
Hello, world
Starting this site as a place to put my projects and writing.