White House AI Czar Sacks Says 'Evidence' DeepSeek Leaned on OpenAI's Models | Hacker News Reader

White House AI Czar Sacks Says 'Evidence' DeepSeek Leaned on OpenAI's Models

bloomberg.com

27 pointshealsdata17 comments

> DeepSeek leaned on the output of OpenAI’s models to help develop its own technology.

closedai's boss keeps saying it's for the "betterment of humanity" so he's either lying or chinese aren't people. Stealing from the entire internet didn't seem to bother him back then, weird

> Stealing from the entire internet didn't seem to bother him back then, weird

You no underestand: me stealing: good, you stealing: bad. Me (Microsoft, parent of OpenAI): copyright infrigement good, you copyright infrigement: BSA, MPAA, RIAA.

Isn't that lovely ? /s

I think it would be impossible to not use output from OpenAI's models, since a significant fraction of new internet content is AI-generated. They are suffering from their own success.

Oh no, One of the companies that pilfered the world’s knowledge to create its models had its knowledge pilfered.

Anyway…

I am extremely amused at the idea of large language model creators getting huffy about IP theft.

OpenAI is laundering all published works and possible non-public as well, but they are "investigating whether data output from OpenAI’s technology was obtained in an unauthorised manner".

Investigate at home first:

https://sfstandard.com/2024/06/13/telemedicine-adderall-vyva...

https://www.dlnews.com/articles/regulation/buterin-warren-jo...

Really, who cares?

OpenAI has made the argument in the past that it couldn't exist without utilizing copyrighted works without permission and, as a net good for humanity, they should get a pass. Old dusty laws and authors rights should clearly take a backseat to innovation!

... except now the shoe is on the other foot where suddenly it's bad to use things without permission. If you were a proponent of AI and AI accelerationism then this should still be kosher since the "net good" has increased with a completely open source model that is apparently considerably more efficient.

The only remaining line of defense amounts to censorship concerns and/or "China bad," as if OpenAI also doesn't do the same and the US was a geopolitical darling.

If DeepSeek were petty they would use quotes from OpenAI as their justification.

They should train an LLM on their justifications and answer via a chatbot.

We are in a state of denial.

If AI is the new electricity this idea the US is going to monopolize electricity is just dumb.

David Sacks I am sure has many bad bets right now so has to talk his book.

It’s even worse because OpenAI took all that information and put it behind a paywall they want to make tons of money with.

DeepSeek open sourced it so all of humanity can actually benefit from it.

after the deep seek release, banning open source models was an obvious outcome. you better git clone all the open source models but be careful bc microsoft will have that action in a log somewhere

Microsoft Probing If DeepSeek-Linked Group Improperly Obtained OpenAI Data

https://news.ycombinator.com/item?id=42861150

Want to protect your investment bubble and moat? Simply steal. Steal TikTok.Steal DeepSeek. Steal boeing a few more years. Choose Life.

Maybe the upper crust of the oligarchy and its beloved caste of toolsmiths are a little inbred and weak when actual competition rears its head. Almost surprised by the free market.

When free market proponents talk about the free market, they really just mean "our market".

The down round that OpenAI will have will hurt...

Reply on news.ycombinator.com