@CelebratedWhale How will you handle something like “Sora-Preview” or “Sora-Turbo”? Or some other such version which is somehow paired back or restricted compared to the full Sora model?

@LiamZ I will resolve that in the affirmative if it has Sora in the name and is a text to video model

Can we resolve Feb, Mar, and Apr to NO?

bumping this, is there a reason why the month that past are not already resolved to NO?

Yes, the reason is that linked multiple choice markets cannot resolve some options early. we're going to have to wait for sora to be available before any options can resolve

oh i see

If they announce it will never be publicly available, does this resolve to Other? @CelebratedWhale

@Bayesian Yes

Do all get resolved to yes ahead of the time of release? Like if it is released in July, would december resolve YES?

@Cadberry No, it resolves to the time it first becomes publicly available

Arbitrage opportunity:

I do not think there is enough GPU capabilities to run it in production at scale. It will take over a year before it is released imho.

There is no chance this releases before the US election

1. GIANT election interference risk. Most people aren't even 'aware' that AI can do art or voices yet, and most haven't even used ChatGPT yet. Imagine SORA being unleashed.

2. Giant computation demand. SORA looks extremely intensive to run, so will eat up GPUs for GPT-5 training.

If you have to request access and it's not granted automatically it doesn't count?

@nikstar Correct