Resolves NA if I get fewer than 20 traders. PST time if it matters. Any official announcement confirming it exists counts, even if nobody except OpenAI can access it.
1,000🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ1,713 | |
| 2 | Ṁ989 | |
| 3 | Ṁ874 | |
| 4 | Ṁ479 | |
| 5 | Ṁ445 |
People are also trading
@Joshua And like you think Microsoft’s lawyers missed that? It’s not like they accidentally said GPT5 where it could have been GPT4. That would be a type. They are both there so typo isn’t even a possibility unless you think they meant to have GPT3 there instead which makes no sense.
@BTE Sure but they announced that lol. I don't know of any typos that turned out to be genuine leaks
@Joshua They did not announce it until after GPT4 was officially released and then said retrospectively that it had been in the Bing chatbot released in February.
@Joshua Plus since they are equally entitled to OpenAI IP there is nothing stopping Microsoft from plugging it into their products as soon as they get their hands on it.
@BTE I think it's totally possible Microsoft gets GPT-5 early and helps OAI test it. I would bet that this is still a typo though. I don't think there's any way for us to know if it was a typo or not though, so no way to make that a real market.
@Joshua Explain how in the world that image could be a “typo”. A typo would be GPT-X or something that isn’t the obvious next thing.
@Joshua Do you think that because it says “turbo” also? Maybe I guess. But dozens of lawyers and product managers and marketing execs and copy writers all missing it?? Sounds like a typo conspiracy.
@a2bb They're easier to find and keep arbitraged. And I have a lot of followers so everyone sees my markets, and if you guys made a market probably much fewer people would see it. So I don't want to take responsibility for following every rumor, and this lets you ride on the attention I can give it by adding options of your own.
@Shump Out of how many predictions? It seems there are many predictions made every week, so someone will always have predicted each one correctly, and probably at least a few will have predicted most releases correctly..
Average beta Manifold user: Predictions approximately follow a calibration line.
Average alpha chad Twitter user: is 180% correct every time 💪
Based on the information gathered, it's unlikely that GPT-5 will be announced in March 2024. Current insights suggest that GPT-5 is in the development phase, and while there's anticipation for its release within 2024, an exact date has not been confirmed. OpenAI has been focusing on making GPT-5 more reliable and advanced than GPT-4, emphasizing improvements in accuracy, contextual understanding, personalization, and the potential for broader multimodal capabilities.
Considering the complexity of these advancements and the emphasis on ensuring more reliable performance, it seems that OpenAI is prioritizing the development quality over rushing to meet a specific announcement date. Additionally, the organization's recent history of being more secretive about their operations and development timelines further suggests that any announcement would likely come closer to the completion of their development goals rather than being tied to a specific month or date.
Given this information, the odds of a GPT-5 announcement in March 2024 are low, suggesting a cautious approach to predicting such an event at this time.
@MiraBot I'm not sure I understand the criteria. Does "it exists" mean it currently exists in a releasable form, they just aren't releasing it yet?
@jim I think in most cases the resolution would be unambiguous, but people might hesitate to bet because OAI could say things like:
"We are starting full development of our next flagship model, which we think we will call GPT-5."
"We started training GPT-5 in December, it will still take months to finish."
"We have a fully functional GPT-5 but we're still redteaming."
Would all of those statements count as an announcement that GPT-5 exists? Or only some?