On a more serious note, every time I see a report like this from you, I see the headline and I think, no, please not again. But it doesn't surprise. I've taken part in environmental conferences where attending trade ministers surreptitiously edited the language of official proceedings to water down things like the precautionary principle... gee that sounds awfully similar to this AI regulation... it's the same game, played on different fields.
Just on the issue itself - there isn’t any *actual text* yet, only political agreements that are currently interpreted by people writing said text. Which looks like it will take another month.
Some water to your mill Gary, take a look at the comments here about people doubting the authenticity of a live chat between Elon Musk and a few other people :
People (me included) are getting increasingly doubtful as to whether or not audio conversations posted online are real or not. And that's just the beginning, imagine having to doubt any video involving celebrities ? What happens when AI gets so good you can fake a video of your neighbor doing something illegal from just a few old videos of him at a party ? There's a really nasty roller coaster track coming ahead.
This is pretty straightforward. AI safety threatens their business model, market share, their wealth! They believe. Therefore, the big tech aristocracy, like aristocracies of the past, will lie and cheat to protect their enormous wealth generating ecosystem.
Hanlon's Razor, maybe? I'm not that convinced these 'tinkering engineers trying to get beyond light speed' are all that smart and he might simply not have been informed. Besides, is it possible the proposal was old, but its addition to the draft text was late? Just wondering, because in general stupidity is a better explanation than evil.
Aren’t these the facts though? Meta’s AI Chief publicly claims foundation models were added late. They weren’t. And it was the French government who pushed for it in the first place, more than a year ago.
"Gary Marcus has seen Yann LeCun lie before, and documented it, but this takes the cake."
It's much worse Gary. This takes the chocolate. It takes the world's supply of Vosges! 😱
On a more serious note, every time I see a report like this from you, I see the headline and I think, no, please not again. But it doesn't surprise. I've taken part in environmental conferences where attending trade ministers surreptitiously edited the language of official proceedings to water down things like the precautionary principle... gee that sounds awfully similar to this AI regulation... it's the same game, played on different fields.
Just on the issue itself - there isn’t any *actual text* yet, only political agreements that are currently interpreted by people writing said text. Which looks like it will take another month.
in the best case, perhaps
Please join Threads, Mr. Marcus! Europe is in now. Let's go!
Some water to your mill Gary, take a look at the comments here about people doubting the authenticity of a live chat between Elon Musk and a few other people :
https://www.youtube.com/watch?v=yCkQH1T-X6A
or
https://www.youtube.com/watch?v=BEW2aYQlXo4
People (me included) are getting increasingly doubtful as to whether or not audio conversations posted online are real or not. And that's just the beginning, imagine having to doubt any video involving celebrities ? What happens when AI gets so good you can fake a video of your neighbor doing something illegal from just a few old videos of him at a party ? There's a really nasty roller coaster track coming ahead.
This is pretty straightforward. AI safety threatens their business model, market share, their wealth! They believe. Therefore, the big tech aristocracy, like aristocracies of the past, will lie and cheat to protect their enormous wealth generating ecosystem.
Throwing shade... the oldest trick in the book!
Hanlon's Razor, maybe? I'm not that convinced these 'tinkering engineers trying to get beyond light speed' are all that smart and he might simply not have been informed. Besides, is it possible the proposal was old, but its addition to the draft text was late? Just wondering, because in general stupidity is a better explanation than evil.
i can’t speak for 2022, but people worked on it more or less continuously throughout 2023
But it was too early then to be regulating something that nobody understood what it even was!
Doesn't seem too early to have been considering regulation. Perhaps it was too early to have been establishing regulation.
But I think the point of this post is the unmitigated deception.
But these are the facts. Where is the drama?
Aren’t these the facts though? Meta’s AI Chief publicly claims foundation models were added late. They weren’t. And it was the French government who pushed for it in the first place, more than a year ago.