Tibo runs 10 agents at once and calls his brain fried
Tibo says he context switches 100 times more than he did five years ago, and that almost every founder he talks to says the same. Meanwhile DHH points at a Rails benchmark where the cheap model costs a tenth of the accurate one.
Real money
Tibo says 10 agents at once left him more tired than ever
Tibo says he runs up to 10 agents at the same time most days and context switches roughly 100 times more than he did five years ago. He calls it the fried brain effect, says he has no solution, and says almost every founder he talks to reports the same thing.
we got superpowers, and somehow the pressure only went up. if you can ship 10x more, you feel like you have to ship 10x more
levelsio puts the Nomads data behind an API at $2,490 to $9,990 a year
levelsio launched a paid Nomads API starting at $2,490 a year and going up to $9,990, built primarily for AI agents and secondly for companies integrating the data. It covers cost of living, internet speeds tested by nomads rather than marketed speeds, weather, quality of life scores and reviews from tens of thousands of visitors.
This week we talk a lot about moats, and data is a moat, and Nomads has a lot of data that I can get paid for
DHH: Sol lands within 5 points of Opus at a thirteenth of the cost
The Rails team published Stage 3 of its agent benchmark, seven multi-step tasks spanning three to eight tickets each: GPT-6.1 Sol hit 60.8% success at $0.80 per run, Opus 5.5 at max effort hit 65.3% as the most accurate but also the most expensive and among the slowest.
Sol completes the tasks less than 5 percentage points off Opus, but at less than a tenth of the cost!!