Just dropped a preview of some of the latest work I'm doing @precursorlabs to distill personal judgment into local models continuously.
Our main finding is that models are overall pretty bad at predicting the decisions you actually make on a day to day basis.
If you had a local "decision model" that was able to actually make the decision you would, given the same context, you could move 1000x faster.
There seems to be consistent learning as a function of data, pointing to something akin to local scaling laws!
@vixalientweets For the sake of everyone else: addressed this in Discord (https://t.co/7jojQotMS2) and in https://t.co/95ahHaLXBA
Thanks for bringing this up!
$3/week for GLM and DeepSeek API access
Pareto Pass is officially live 🎉
70 tok/s median on GLM-5.3
Join with a referral: $1 for your first week
Refer a friend: earn $3 subscription credit after their first $3 renewal
Get started: https://t.co/xV8Ecip925
After digging into how inference providers cut corners, I tested GLM 5.2 on FP4 endpoints. They charge less per token and more per answer.
A listed 32% discount became an 11% premium per solved problem. Price parity became a 40% premium.
Have a read!
We're a small team of engineers, financiers, and company builders working across three threads: multi-agent systems, the AI buildout and market structure, and decision-making infrastructure.
We'll be sharing research and essays here and on our Substack (https://t.co/re7AEPYfbD).
Follow along!
Introducing Precursor Labs: a research company studying the infrastructure for collective intelligence.
The next generation of autonomous systems will need to coordinate, reason under uncertainty, and allocate resources in dynamic environments. We're working on how. ↓