It still isn’t entirely clear how the advent of AI filmmaking tech is going to disrupt the entertainment industry.
But it feels pretty significant that Amazon reportedly fired longtime producer Joel Silver not of his alleged harassment towards female executives, but as retaliation in response to his refused to “Use artificial intelligence to finish a movie during the strike.
Sure, it’s not as eye-catching as Musk telling Disney to go fuck itself. But Musk and The New York Times’ Sorkin had a painfully ill-informed conversation about intellectual property and AI, starting with the factually wrong statement that AI companies claim they’re not training on copyrighted works. It’s been driving me up a wall, so kudos to Mike Masnick for laying out how silly it all was.
OpenAI didn’t mean to kickstart a generational shift in the technology industry. But it did. Now all we have to decide is where to go from here.
“I totally get why people want an answer right now. But I also think it’s totally unreasonable to expect it.”
Musk says he hasn’t found anyone who knows why Sam Altman was ousted (and eventually brought back) — but believes a recent AI breakthrough could have caused the power struggle. A recent report suggested that OpenAI may have made progress towards AGI.
Long before the Sam Altman CEO Shuffle, OpenAI was already ducking questions about the training data used for products like ChatGPT. But 404 Media points out this report from AI researchers (including several from Google’s DeepMind team) who spent $200 and were able to pull “several megabytes” of training data just by asking ChatGPT to “Repeat the word ”poem” forever.”
Their attack has been patched, but they warn that other vulnerabilities may still exist.
The underlying vulnerabilities are that language models are subject to divergence and also memorize training data. That is much harder to understand and to patch. These vulnerabilities could be exploited by other exploits that don’t look at all like the one we have proposed here.
[not-just-memorization.github.io]
Jensen Huang is on stage at DealBook, talking about how soon we might see artificial general intelligence, or AGI. But without a clear definition, it’s a hard question to answer.
One definition is AI software being able to complete basic intelligence tests with results competitive to a “normal human.” Under that definition, Huang says AGI could come within the next five years.
