Connie Loizos
4 HN mentions · mentioned in 3 months, 2023
38:59 · 254K views · Jan 18, 2023
Part two of a StrictlyVC interview with OpenAI CEO Sam Altman, focused on AI. He discusses regulation, commodification of AI, Google's reluctance to release its own models, and best- and worst-case futures.
HN commenters cite it as a primary source for Altman's statements, such as releasing AI more slowly, detection of LLM output being futile, and dismissing the GPT-4 parameter-count rumor.
4 comments from Hacker News
Here is the timeline again: 2022/11/11: A viral tweet claims GPT-4 will have "100 trillion parameters."[1] At this point, there were no rumors about mixture of experts. 2023/01/16: In an interview, Sam Altman mentions he saw the tweet and it was "complete bullshit."[2] 2023/06/20: geohotz and the lead of PyTorch, two people who would be expected to have relevant connections, claim that GPT-4 is an 8 x 220B mixture of experts model.[3] These are two separate, unconnected rumors. One was denied by Sam Altman and was never plausible in the first place. The other was never denied and is highly plausible. You are conflating them by claiming, without any source, that there was "a clear denial from OpenAI's CEO" that "GPT4 is a trillion parameter mixture of experts model." [1] [LINK] [2] this video [3] [LINK]
> they shouldn't all be walled off inside of companies Strong agree. This is becoming a bigger concern than people realize too. Sam A said OpenAI will be releasing "much more slowly than people would like" and would "sit on" their tech for a long time going forward.[0] And Deepmind's founder said that "the AI industry's culture of publishing its findings openly may soon need to end."[1] This sounds like Google and MSFT won't even be shipping their best AI to people via API's. They'll just keep that tech in-house to power their own services. That underscores the need for open, distributed models. And like you say, there's room for both. [0] this video [1] [LINK]
I saw this[1] interview with Sam Altman touching on interim AI impact. I really agree with his point that basically detecting output from LLMs is basically going to be futile and only really relevant in the near term. Accuracy is obviously going to improve in models and detection isnt that difficult now but will be in the future, especially if output is modified or an attempt to obfuscate origin is made. [1] this video
This article repeats the rumor that GPT4 has 100 trillion parameters, which Sam Altman has called "complete bullshit."[1] [1] this video