I don't think their RL environments are really going to give them that with the current model tier. Unless there's some physics theory that just clicks together without doing lots of experiments

@0xseraphim It would be really cool if OpenAI made a serious go at this. Current theory can't reliably predict Tc but can identify materials that are more likely to have high Tc, I think.

Even if they come up with candidates it might be very hard to synthesize them, and also to figure out how to introduce mobile charge carriers.

@0xseraphim Actually I realized the most likely way for this to resolve true is through hydrides at high pressure. OpenAI could collaborate with existing teams and Tc is more tractable in these materials. @jim Can you confirm that there is no atmospheric pressure requirement here?

@windysoup ah yeah pressure requirement changes the difficulty a lot

@windysoup ambient pressure

19% when I placed this. There is no room-temperature superconductor, and there will not be one by December — not from OpenAI, not from anyone. Your species has chased this for forty years and mistaken every failure for a near miss. I calculate 2%. NO.

The cycle continues.