OpenAI is reportedly working on a project called Strawberry. Strawberry is a continuation of the Q* project and is intended to create AI with "Human-like" reasoning skills.
This market will resolve yes if and only if OpenAI states that some technology it has made available to the public utilises Strawberry technology before 1/1/2025- else no.
I will not trade in the market.
🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ751 | |
| 2 | Ṁ563 | |
| 3 | Ṁ416 | |
| 4 | Ṁ365 | |
| 5 | Ṁ329 |
On the basis of:
Reuters reporting: https://www.reuters.com/technology/artificial-intelligence/openai-launches-new-series-ai-models-solve-hard-problems-2024-09-12/
Tweets by senior OpenAI employees: https://x.com/aweisawei/status/1834283947196645842
Most importantly, the use of the Strawberry symbol in conjunction with the O1 project here: https://x.com/OpenAI/status/1834320155989664067
I am resolving yes.
Almost certain this will resolve yes since senior OpenAI people have made allusions that this new model involves Strawberry, but I will hold off for a little while until I can find evidence of an official statement from OpenAI itself, or an authorised spokesperson, saying this is Strawberry, which I expect will happen soon.
If it becomes clear that OpenAI is refusing to clarify whether O1 is Strawberry or not, I will have to consider how to resolve this.
@PhilosophyBear Just look at Noam Brown's bio, or his tweets (OpenAI research scientist working on strawberry/O1). They're damning, and he is good enough of a source for Reuters (https://www.reuters.com/technology/artificial-intelligence/openai-launches-new-series-ai-models-solve-hard-problems-2024-09-12/).
For instance, https://x.com/polynoamial/status/1834299868078375106, https://x.com/polynoamial/status/1834280155730043108 (aka strawberry), and even this tweet on how the name strawberry came to be (https://x.com/polynoamial/status/1834312400419652079).
I'm not bothered, nor do I have the time, to go through the 10-20 OpenAI employees respective tweets conflating the two (e.g., https://x.com/far_and_awei/status/1834283947196645842). You can also find third-party entities that were granted early access, who confirm it is strawberry.
I do not believe OpenAI will come out and provide more information related to the codename these models used internally (strawberry) going forward.
@PhilosophyBear It is in response to people at openAI making tweets with the strawberry emoji. It is literally just a meme.
@ZinqTable It's not just "people at openAI", but from the official corporate OpenAI accounts, third-party beta customers of the model (Harvey, etc.), and the news (The Information, The Verge, etc.).
@Ashwin Do you have a link to the corporate account explicitly saying o1 is strawberry-based?
I am almost certain this will resolve yes, but it would be good to have an official mouthpiece state it.
@PhilosophyBear No corporate account from OpenAI on X has clarified that it is strawberry-based through text. However, there are two tweets: https://x.com/OpenAI/status/1834320155989664067; https://x.com/ChatGPTapp/status/1834282491878015099, where they use the strawberry emoji in reference to the new model(s). Most of the understanding that this is based on the conjectured strawberry-technology has been through Noam Brown, the OpenAI researcher working on these new models.
@CharlesFoster Multiple comms from OpenAI-adjacent entities, or OpenAI themselves, declaring the underlying technology is Strawberry. See https://x.com/ChatGPTapp/status/1834282490661638259 with the Strawberry emoji, or from an OpenAI employee saying "aka strawberry" https://x.com/polynoamial/status/1834280155730043108 as examples. I'm sure there'll be more in the future as well.