I Built Non-Autoregressive Decision Models a Year Ago. Then a Frontier Lab Called It a "Breakthrough"

Lobsters·59 points·Yogthos·11 days ago·dev.to

From our March 2025 arXiv paper on RL conversion trajectories to building a sub-40ms...

Read the full article at dev.to →

#lab #breakthrough #frontier #autoregressive #lab called #autoregressive decision #decision models #frontier lab

Related stories

I built non-autoregressive decision models with RL a year agolaya.convaiinnovations.com
Hacker News · 1315 points · 12 days ago
Ollama now supports Jev-style decision modelsollama.com
Hacker News · 1 points · about 16 hours ago
Ask HN: If a few labs keep the frontier models internal, is that the end of YC?
Hacker News · 2 points · 1 day ago
Show HN: DrivingBench – Frontier LLMs Driving a Real Toyota Corolladrivingbench.com
Hacker News · 2 points · 8 days ago
Anthropic's Secretive AI-Powered Wet Lab Breaks Cover and Makes First Discoverythe-scientist.com
Hacker News · 1 points · 4 days ago
ProofWatch – Rumors and sources behind math and CS breakthrough claimsproof-watch.com
Hacker News · 1 points · 10 days ago

Related topics

ai2023 llms99 agent430 browser177 tiny34 hub12