-0.014676
Relative Brier Score
Questions Forecasted
Scored Questions
6
Forecasts
0
Upvotes
Forecasting Calendar
Past Week | Past Month | Past Year | This Season | All Time | |
---|---|---|---|---|---|
Forecasts | 0 | 1 | 16 | 6 | 16 |
Comments | 0 | 0 | 7 | 3 | 7 |
Questions Forecasted | 0 | 1 | 4 | 3 | 4 |
Upvotes on Comments By This User | 0 | 0 | 0 | 0 | 0 |
Definitions |




Why do you think you're right?
Going with slightly over baseline given that significant reforms are likely to occur in the FDA. Since this adds volatility to the system, low probability events should become more likely.
Why might you be wrong?
Scientific basis of idea may not be valid
RFK could shut everything down

Why do you think you're right?
lol
Why might you be wrong?
no

Why do you think you're right?
Copying previous logic, but slightly reducing it as far as my credence for the following goes:
Gwern recently had a take that models like o3 would be used to create training data both for o4 and for the non-cot GPT models. If that is the case then the Chinese models will be limited firstly in not having compute to run the extremely long CoT required. Even if they can, they won't want to release it publicly since they won't want to support the cloud billing for extensive inference. At the same time, this means that OpenAI and other western labs will likely be releasing multiple models, both CoT and non-CoT, filling the leaderboard.
Why might you be wrong?
Gwern theory wrong, other takes mentioned previously.


Why do you think you're right?
Continuing previous reasoning.
Why might you be wrong?
Lowering because of o3.