Twitxy Saved

David Minarsch ☴

Verified account Olas (formerly Autonolas)

@david_enim

Co-founder at @valoryag + builder at @autonolas + chief vibe coder @pearldotyou + creator of @aea_dev | Pushing the autonomous edge
London davidminarsch.github.io Joined June 2011
1.6K Following 4K Followers 4.1K Posts

Latest posts 5

Pinned post

David Minarsch ☴ Verified account @david_enim
Bitcoin: decentralised. one app.

Ethereum: decentralised. infinite apps.

[?]: decentralised. n# apps.

The design space of an app chain that tightly controls (via onchain governance ideally) what apps (smart contracts) can be deployed but always keeps permissionlessness and censorship resistance at the individual transaction layer is definitely under explored.

Then all the apps can be precompiles, and everything onchain and offchain (indexers) optimised for the exact apps deployed.
3 replies 2 reposts 15 likes 1.8K views View on X
David Minarsch ☴ Verified account @david_enim
2nd open-weight fine-tune from Olas ecosystem that performs competitively with closed AI on forecasting in prediction markets context.

We'll keep chipping away at this.

Check out the blog and model!
Olas (formerly Autonolas) Verified account @autonolas Nov 11
A lean, open-weight model that comes within 1.2 points of GPT-5.6 at forecasting real-world events.

Core contributor @valoryag trained Olas-Predict-Qwen3.8-27B, an open-weight forecasting model built on Qwen3.8-27B, using reviewed forecasting examples and reinforcement learning. It follows Olas-Predict-R1-14B, Valory’s first Olas Predict model.

In Valory’s evaluation on 6,824 prediction markets created after GPT-5.6’s training cutoff, with the same prompts and evidence for both models:

• Balanced accuracy (higher is better): 67.3% vs 68.5% for GPT-5.6
• Brier error (lower is better): 0.2124 vs 0.2029 for GPT-5.6
• Against Olas-Predict-R1-14B: 22% lower Brier error

It gets this close from a starting point that scores 15.6% on the AA-Omniscience factual-knowledge benchmark, against about 59% for GPT-5.6.

Most of the gain came from the stronger starting model. Valory’s domain training added 5% lower Brier error and 12% lower calibration error on top.

Next, Valory aims to close the remaining gap in future training experiments.

Read the results: olas.network/blog/meet-olas…
View quoted post on X
2 replies 1 reposts 19 likes 466 views View on X
David Minarsch ☴ Verified account @david_enim
When doing research negative results are also valuable.

Team's sharing one here.

If you're into this sort of work and would like to join the @valoryag team, DM me.
Olas (formerly Autonolas) Verified account @autonolas Nov 10
Bigger wasn’t better this time.

Core contributor @valoryag fine-tuned NVIDIA’s Nemotron 120B on forecasting data from Olas agents, using the recipe behind Olas-Predict-R1-14B: selected reasoning examples, a supervised warm-up, then reinforcement learning. Training ran through @thinkymachines’ Tinker.

On 11,496 held-out predictions, Nemotron showed no clear improvement. Before → after training:

• Brier error (lower is better): 0.265 → 0.267
• Balanced accuracy (higher is better): 59.7% → 59.4%
• Unreadable answers (fewer is better): 503 → 528

Our trained 14B model had better calibration, higher balanced accuracy and just 10 unreadable answers. Nemotron retained a slightly lower Brier error: 0.267 versus 0.270.

Why didn’t the recipe transfer? Nemotron’s mixture-of-experts design may need more varied data or different training settings for forecasting. We haven’t tested those explanations, so the cause remains open.

What we learned: our existing recipe and training budget improved the 14B model but brought no clear gain to Nemotron.

Next, we’ll test a larger dense model, with an early check for improvement before committing to the full run.

Read the experiment and results: olas.network/blog/nemotron-…
View quoted post on X
1 replies 1 reposts 10 likes 425 views View on X
David Minarsch ☴ Verified account @david_enim
Interesting how a proposed updated to Ethereum is once again testing developers and the notion of immutable apps.

Maybe it’s time to build the third way?

Getting increasingly bearish on Ethereum - death by thousand cuts (to all developer and ETH holders). Maybe I’m too pessimistic?

x.com/z0r0zzz/status…
David Minarsch ☴ Verified account @david_enim Aug 19
Bitcoin: decentralised. one app.

Ethereum: decentralised. infinite apps.

[?]: decentralised. n# apps.

The design space of an app chain that tightly controls (via onchain governance ideally) what apps (smart contracts) can be deployed but always keeps permissionlessness and censorship resistance at the individual transaction layer is definitely under explored.

Then all the apps can be precompiles, and everything onchain and offchain (indexers) optimised for the exact apps deployed.
View quoted post on X