@rosstaylor90 General Reasoning, the Hedgefund?
@xeophon Yep, on KellyBench alone the difference from frontier is narrower. On our internal quant evals it looks a little worse - hoping we’ll be able to publish on those soon 🤞.
A quick social media back-and-forth turned the spotlight on General Reasoning after Florian Brand floated a hedge-fund comparison tied to the company's KellyBench benchmark, only for CEO Ross Taylor to reply with a shushing emoji and zero added context.
@rosstaylor90 General Reasoning, the Hedgefund?
@xeophon Yep, on KellyBench alone the difference from frontier is narrower. On our internal quant evals it looks a little worse - hoping we’ll be able to publish on those soon 🤞.
KellyBench's simulated betting-market setup makes the analogy easy to reach, yet the company remains described only as an AI research startup with no evidence of hedge-fund activity.
Neither participant shared technical or organizational information, so any link between the shush and sensitive plans stays speculative.
No Digg Deeper questions have been answered for this story yet.
@xeophon 🤫
@rosstaylor90 General Reasoning, the Hedgefund?