3 ms·
Ask HN: If Sam Altman succeeds, people will just use Chinese/Russian AI instead
I mean, if the regulations require a special government license to develop large language models, then that will stop US universities and open source community from making models competitive to GPT-4, 5, etc. But nothing stops people from using similar models built by China or Russia.
- david927 3y agoYes. This is just anti-competitive behavior and it won't work.
- _the_inflator 3y agoThe company that named itself "Open". ;)
- smoldesu 3y agoIf the handcuffs are golden enough, US regulators will gladly try them on.
- wmf 3y agoSupposedly LLMs are already banned in China for being counterrevolutionary and don't make me laugh about Russia. Maybe we'll see an AI lab buy up FTX's offices though.
- PaulHoule 3y agoIn China an LLM would have to balk at requests like "Give ten reasons why Taiwan should be an independent country" or "What happened on June 4, 1989?". Oddly, most search engines give this answer https://en.wikipedia.org/wiki/1989_Tiananmen_Square_protests_and_massacre https://en.wikipedia.org/wiki/1989_Tiananmen_Square_protests... but a demo of the Alpaca model gave a different answer. (... hmm, Toronto Blue Jays beat the Red Sox) If I added "in china" to that question it did give that answer but added something to the effect of "... but they don't like to talk about it".
- scrum-treats 3y ago"But nothing stops people from using similar models built by China or Russia." It's not a mystery that US is in cold war with Russia and China. Nor is it a mystery that there is advanced AI beyond GPT-4 that we poor public don't have access to. But the government does. The idea would be to establish rules for public vs. private use, build sophisticated detection mechanisms presumably in coordination with the rest of the world (because world domination is a thing), and cripple any competing advancements. We can say it's anti-competitive, reduce the scope of awareness, however it's shoring up to be a significant concern in terms of national security. And humanity. Interested to see where this goes, and how fast...
- sharemywin 3y agoChinese/Russian AI, aren't there distributed decentralized versions already?