As most of you already know openai released a new model called o1 capable of reasoning. In OpenAI's words the model can:
spend more time thinking before they respond. They can reason through complex tasks and solve harder problems than previous models in science, coding, and math.
I wonder if Anthropic is already working on something that gets released this year. Even if it’s invite-only, we need to see evidence that the model exists and that someone outside Anthropic’s network of family and friends has access to it. The model should be able to reason for different timeframes based on the complexity of a problem before coming up with a response.
/Soli/will-a-finetuned-opensource-model-u
1,000
3.00@JoshYou good question, by this logic gpt-4o was already doing some reasoning/thinking but this market is clearly looking for something much more significant. i expect anthropic to make a press release like they usually do that contains the words chain-of-thought but would not require this as a condition. let’s just say the model should be able to reason for different timeframes same as o1 based on the complexity of a problem
@JamesBaker3 you are assuming they haven't been working on a similar model already which i think is an incorrect assumption