Nvidia predicts AI models 1,000,000X more powerful than ChatGPT within 10 years

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by epups on Saturday February 25, 2023 @02:58PM (#34939859)

There is not enough written text to scale a model like ChatGPT at even 10x its current parameters, so improvements will have to come from optimizations and/or longer training (at diminishing returns).

[ Parent | on HN ]

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by motoxpro on Saturday February 25, 2023 @03:05PM (#34939922)

Enough data wont really be a problem. https://us.sganalytics.com/blog/2-5-quintillion-bytes-of-dat...

I wouldn't be surprised if we create more data in one year than is in the entire internet from the start

[ Parent | on HN ]

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by KeplerBoy on Saturday February 25, 2023 @03:12PM (#34939982)

quintillion bytes of social media probably have very limited value.

[ Parent | on HN ]

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by motoxpro on Saturday February 25, 2023 @03:16PM (#34940009)

Really? All of humanity imparting everything they think and know publicly seems incredibly valuable. 100% serious.

[ Parent | on HN ]

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by BooneJS on Saturday February 25, 2023 @03:49PM (#34940314)

If you're looking to make a tweet generator, sure. If you're looking to harness the power of facts, no.

[ Parent | on HN ]

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by rvnx on Saturday February 25, 2023 @07:27PM (#34941886)

The main issue might be that the training set may get polluted by a lot of AI generated text (content that is not supposed to be there, but is there because AI-generated content is becoming widespread online).

[ Parent | on HN ]

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by zamnos on Saturday February 25, 2023 @10:45PM (#34942848)

Storage is cheap. With ChatGPT being the preminent system, and with OpenAI recording all of its output, it should be possible to exclude modified answers from the training data.

[ Parent | on HN ]

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by motoxpro on Saturday February 25, 2023 @09:48PM (#34942568)

So interesting that people think facts are the only thing that's valuable. Your comment isn't a fact, does that mean it's worthless?

[ Parent | on HN ]

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by ranguna on Sunday February 26, 2023 @03:32AM (#34944013)

When looked at individually it's not worthless, but when you are talking about a corpus that significantly lacks facts, then the corpus itself is worthless for any mainstream usage other than a social network text generator.

[ Parent | on HN ]

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by ThrowawayTestr on Saturday February 25, 2023 @05:58PM (#34941322)

How do you know that? There are thousands upon thousands of books that haven't been digitized.

[ Parent | on HN ]

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by ninethirty on Sunday February 26, 2023 @02:15AM (#34943740)

GPT-4 Will Have 100 Trillion Parameters — 500x the Size of GPT-3

[ Parent | on HN ]

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by ETH_start on Sunday February 26, 2023 @04:31AM (#34944260)

That's likely a massive over-estimation. I would bet it's around 1 trillion.

[ Parent | on HN ]

Re:Nvidia predicts AI models 1,000,000X more powerful than C... by imtringued on Sunday February 26, 2023 @01:37PM (#34948046)

That is just an internet rumor and officially denied by OpenAI.

[ Parent | on HN ]