Eric Cho

building Trata (YC W25), hedge-bench

[Eric Cho @_ericcho]

The coolest publicity we could ask for. I’ve been a fan of Money Stuff for years!!!

@matt_levine 🤝 @trytrata

Public investors: My DMs are open. Funding secured. As always feel free to reach out anonymously. Acceleration coming soon

https://t.co/I6GJLee8sD

[Eric Cho @_ericcho]

Yeah!!

[Trata (YC W25) @trytrata]

Today, Trata is excited to share the revised Hedge-Bench: our benchmark tying open-ending reasoning to prediction capabilities.

We constructed 92 real-world environments derived from hedge fund analysts’ reasoning traces. Agents reason from frozen information snapshots, and their forecasts are evaluated against future real-world outcomes. No model scores above 50% (GPT-6 Astra being the top performer at 44.2%). For frontier, state-of-the-art models, these results highlight a persistent and wide gap between producing plausible-sounding reasoning and making accurate, economically valuable forecasts. In a semi-verifiable domain like finance, where outcomes are delayed and reasoning is difficult to verify, intermediate expert reasoning traces are crucial.

We post-trained @Meta's Muse-Glimmer-30B on 165 environments with per-turn credit assignment using an LLM-as-a-Judge. We observe evidence of generalization: held-out forecasting accuracy at high effort improved by 10.06 percentage points, while out-of-distribution evaluation scores improved by 3.94 points on @OpenAI's GDPval, 2.29 on @Surge’s GDP.pdf, 7.07 on @Harvey’s LAB, and 4.96 on @Samaya_AI’s FrontierFinance under their respective metrics.

[Christian Keil @pronounced_kyle]

nothing like absolutely hammering out some slides from the middle seat in economy

[Eric Cho @_ericcho]

@YetAnotherValue Legendary

[Eric Cho @_ericcho]

@abundant_labs @baseten Awesome

[Wes Winder @weswinder]

tokenmaxxing is a disease

95% of you are building nonsense

[Eric Cho @_ericcho]

@gregpr07 @mamagnus00 data industry in one sentence

[mark @mmarkgm]

how do you become an arms dealer? do you just wake up and start cold emailing dictators?

[Eric Cho @_ericcho]

@joshearle Wait a minute I didn’t take this photo lol

[Eric Cho @_ericcho]

@joshearle This one was tough

[Eric Cho @_ericcho]

@syriansigma Because @adambcohen93 bought it with his private equity

[Eric Cho @_ericcho]

@syriansigma Even the syrian sigma is mortal

[Eric Cho @_ericcho]

@syriansigma @adambcohen93 @Standard_Cap Hahahaha

[Eric Cho @_ericcho]

@adambcohen93 @Standard_Cap so well deserved, amazing to see 😍😍😍

[Kaya Jones @kayacancode]

What @sama hated most about running yc lol

[Eric Cho @_ericcho]

@adambcohen93 weave your way through

[Eric Cho @_ericcho]

@_WEEXIAO Amazing

[Trata (YC W25) @trytrata]

Fable 5 (@AnthropicAI) scores 22% and tops the Hedge-Bench leaderboard. Running Fable was roughly 2X more expensive than Opus 4.8 per trial. For an industry where accuracy is mission critical, human judgement isn't going away

[Eric Cho @_ericcho]

@syriansigma Spoken like a true sigma

[Eric Cho @_ericcho]

@atelicinvest lolz

[Eric Cho @_ericcho]

@syriansigma The Syrian sigma only asks the best questions