According to a “leak,” OpenAI is working on a project codenamed “Strawberry” — a new AI technology that can reason! Totally like a human! [Reuters, archive] “Strawberry” is a new name for the proje…
You’re very correct! Any quote I could have taken from the article was about LLMs, and that’s my entire point. The Strawberry project from open ai specifically uses tools on their LLMs, so it’s weird to me to dismiss the project by pointing out flaws in LLMs that openAI specifically recognizes and is using other methods to overcome (or so they claim)
And “apparently” is a very load-bearing word. I don’t think AI does nearly as much as anyone thinks but I also don’t think it’s been claimed to do very much by the company. I would love to get an article from you about Sammy boy claiming that current AI reasons and does more. All the articles I could find with interviews with him was him talking about the future of AI and usually about AGI.
You’re very correct! Any quote I could have taken from the article was about LLMs, and that’s my entire point. The Strawberry project from open ai specifically uses tools on their LLMs, so it’s weird to me to dismiss the project by pointing out flaws in LLMs that openAI specifically recognizes and is using other methods to overcome (or so they claim)
https://www.reuters.com/technology/artificial-intelligence/openai-working-new-reasoning-technology-under-code-name-strawberry-2024-07-12/
And “apparently” is a very load-bearing word. I don’t think AI does nearly as much as anyone thinks but I also don’t think it’s been claimed to do very much by the company. I would love to get an article from you about Sammy boy claiming that current AI reasons and does more. All the articles I could find with interviews with him was him talking about the future of AI and usually about AGI.