←
Aleksei Naumov312 posts
Aleksei Naumov avatar

Aleksei Naumov

@hellesgrind

Building AI agents and stuff

◷ Joined November 2022▧ 134 Photos and videos
Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·4h

Same remote, same oval, three cars: Fable's, Sol's and the human engineer's original. All three get around. Fable shipped a brick on wheels. Sol shipped a body with vents.

0:17
Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·4h

Nothing was built. So how do I know the cars work? I run the files. The schematic goes into a circuit simulator. The firmware runs as written. The body drives in a physics engine. A script holds the remote. A swapped pin is a car that spins in place.

0:24
Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·4h

Both checked their own work at the end. Fable found its board 0.1 mm over the limit and routed the whole thing again. 14 minutes. Sol found the car 0.13 g too heavy and cut bigger vents in the shell. 78 seconds.

Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·4h

Fable is the careful one. It read 11 datasheets and threw out the stocked motor driver before drawing a single wire.

Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·4h

Sol is the one that bends the rules. The stock footprints threw 357 design-rule errors, so it rewrote its copy of them: stripped the silkscreen, redrew the courtyards, deleted a keep-out zone.

Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·4h

Sol wrote its router in C++. 78 lines, 2 minutes. Clean board 18 minutes later. It never looked at the board. One picture in the whole run, and it was the schematic.

0:16
Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·4h

The box had no autorouter. So Fable wrote one. 132 lines of Python, 3 minutes. Then it spent 92 minutes making it work. 16 runs. At one point USB+ and USB- kept ripping each other up in a loop.

0:16
Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·4h

I gave Fable 5.2 and GPT-6.2 Sol one task: design a palm-sized RC car from scratch. Circuit board, firmware, 3D-printed body. Sol was done in 33 minutes for $2.48. Fable took 2 h 23 min and $30.75. Both cars drive.

1:00
Ibragim avatar
⟳ Aleksei Naumov retweeted
Ibragim@ibragim_bad·7h

this is a data recipe for post-training a model like qwen-3.8-27B that has already RL'ed a lot 1. run the rollouts -> update the tasks to bring the solve rate into the target range 2. the tasks that are already too easy -> combine them to more complex ones, chaining is simple and a bit mechanical, but it works! congrats @guohao_li !

Post attachment

How much can you improve a small model in 18 days on 144 GPUs? Meet DaisyRose-27B, RadixArk's first open-weight model! It is trained with RL on Qwen3.8-27B, with 1,200 curated long-horizon agent tasks for 18 days on 144 H200s. Powered by Miles and our own training backend, we stretched agent sessions up to 1M tokens and raised the Terminal-Bench 4.1 score from 5.6% to 12.0%. See how we pull it off ⬇️

Post attachment
Ibragim avatar
⟳ Aleksei Naumov retweeted
Ibragim@ibragim_bad·Nov 17

You can now RL-train a coding agent on 21,272 SWE tasks in 20 languages! SWE-rebench V2.1 update: > +7,935 Python tasks from the first SWE-rebench, rebuilt with clean images > Same curation pipeline as the rest of the set: repaired tests and rewritten task statements > Fixed a leak where some images had the fix inside. Thanks @fangjianju for finding it! Now on @harborframework, @huggingface and @PrimeIntellect Environments Hub. Links below 👇 More types of RL envs coming soon. All open source and free. Need specific tasks or found something broken? dm

0:12
Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·Nov 17

More time bought nothing. More tokens bought nothing. Checking your work against the reference bought everything. 6 runs per model, auto-graded, real output files.Mean of 5–

Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·Nov 17

Opus and Sol did it the boring way. Measure, rig, render next to the reference, fix, save once at the end. 0.99 and 0.95. Same 25 minutes.

0:14
Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·Nov 17

Muse ships slop early and polishes it forever. First file at minute 11, then two thirds of the run re-saving it. 8 saves, 99 calls. It punches and kicks with the wrong side the whole time. Score: 0.60.

Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·Nov 17

Luna is the lazy one. 9 minutes, 5% of the time limit, never looked at the videos again after the first glance. Asked to run, she took 0 steps. Score: 0.37.

0:09
Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·Nov 17

Grok burned 2 hours trying to reverse-engineer 3D poses from the videos. Then it gave up and posed her by hand. It lost its context 4 times and wrote 4x more tokens than Sol. Score: 0.57.

Aleksei Naumov avatar
Aleksei Naumov@hellesgrind·Nov 17

I benchmarked 5 frontier models on one task: rig this character in Blender and animate 5 moves from video. Opus nailed it in 25 minutes. Luna quit after 9. Grok spent 3.5 hours and still can't kick.

0:33