Nic Moe
Max KennerlyCan't say I understand why, but I'm glad the Chief AI Scientist at Meta is determined to convince everyone that the AI industry does not take trust and safety issues seriously and thus needs to be heavily regulated. 🤷♂️
Laure@maxkennerly some people at #meta are very well intentioned even if they themselves are in the dark about all of the malfeasance at the platform over the last 7ish years.
Femme Malheureuse@LaureM @maxkennerly They have zero excuses for that claimed ignorance after FB had to pay $5 BILLION for its violations of 2011 FTC consent decree. Any corporation taking a hit in the billions of dollars should have educated their employees about avoiding future penalities of this magnitude.
Olav@LaureM what are the odds that if Meta didn't approve of the messaging, he'd be silenced or job hunting
@olavf that too
@LaureM to put it another way: if you, as a company, were spending billions of dollars in dev work on a project, you wouldn't allow the chief architect to publicly go around saying the project is dangerous and needs government intervention... unless there's something in it for you (company).
Simply put, Meta knows AI is going to be regulated, and wants to make sure that regulation maximizes their profits
Simply put, Meta knows AI is going to be regulated, and wants to make sure that regulation maximizes their profits
@maxkennerly As if any sentient human who values their privacy should ever pay any attention to the entities within the corporate holding structure now known as Meta.
Clearly based on FTC consent decree chronologies we haven't regulated social media adequately up to now; AI is yet another opportunity for socmed corporations to screw the public for profits. https://epic.org/facebook-2011-ftc-consent-order/
Mike Hucka@maxkennerly I'm also puzzled (but happy) that it seems these public postings by high-ranking people at Meta are apparently unfiltered by their PR people. I mean, this stuff is gold.
Jorge Caballero, MD@maxkennerly IIRC a few weeks ago there was a leaked internal document from one of the big tech companies that explains the pro-regulatory stance by OpenAI et al. In short, the tools themselves are so powerful yet accessible, that pushing for cumbersome regulatory frameworks is just about the only way they can achieve a competitive moat. IOW: the weekend hacker can run circles around them already, and they're looking to stiffle competition
joel b@maxkennerly what an ass.
andeux@maxkennerly Yann LeCun is amazingly capable at some specific technical tasks, far beyond what most humans could achieve, but output like this makes me skeptical about any claims that he has achieved general intelligence.
Woolfhound@maxkennerly @Finagle_a_Hegel great look when you work for a company with a week-earned reputation for being really unserious about its impacts on the communities within which it operates
@maxkennerly when companies say they want to be regulated, it's because they want a seat near the front of the table; they want the regs to be *their* regs not necessarily the greater good
Nihl L'Amas
Token Sane Person@NIH_LLAMAS @maxkennerly Actually we do know how to make safe & secure *systems* that include software. These days lifts and aircraft both include lots of software.
The problem with voting machines is that it's expensive and not required, so it doesn't happen.
amen zwa, esq.@maxkennerly I expected LeCun, of all people, to be more measured.
VHaschAre they seriously comparing a ballpoint pen to AI? 🤨
ADHDeanASL@VHasch @maxkennerly Perhaps he means a self-writing ball-point pen
J.H.NoyesThey want to establish a monopoly. They're afraid that democratic access to verbose AI is going to do to their station that which their station did to everyone else with egghead AI.
I don't why you'd be happy about that. This is about picking winners and losers, not keeping people safe.
Hughster"We've invented the ballpoint pen"
"Wow! An historic momen… oh wait, when I try to write something, occasionally it writes something completely different?"
"Roll up folks, get your life-changing technology here"
"When I try to write 'Mars is an empty planet', why does it write 'Mars has 3 million inhabitants'? And when I try to write a shopping list, why does it replace it with the ingredients for napalm?"
"We are proud to announce we have supplied our product to every school"
Jaime Robertson@maxkennerly These guys are never as smart/clever as they think they are.
Ash McAllan@maxkennerly the companies that have invested in leading technologies want the government to regulate those technologies to make it harder for new competitors to start up due to the costs of complying with those new regulations, costs which are easy for the leaders to absorb. Its dropping caltrops onto the track behind you.
Duco@maxkennerly this is about meta? I thought it would be about the stupid "don't say gay" and trans hating legislations appearing recently. 🤔
newsorpigalHe has a point.
People who don't understand anything about machine learning are causing nonsensical hysteria, while real problems are not getting any spotlight.
Mike Rodriquez@maxkennerly Wow. I saw that, and thought it was written by some random loon.
Not by someone with actual authority to make things happen in this space.
Now I'm worried!
Rui Seabra@maxkennerly
The calls for regulation are for regulation that ensures new private monopolies where only big companies can win, without fear of some community project eating their breakfast.
@glynmoody
The calls for regulation are for regulation that ensures new private monopolies where only big companies can win, without fear of some community project eating their breakfast.
@glynmoody
Juan@maxkennerly Forgive me for thinking that YLecun was smarter than this. Or more likely he thinks others are dumber than he is. A ballpen and AI being made to sound equivalent? Maybe it makes sense to flat-earthers? A pen has never tried to convince me that it is absolutely right when it makes up information out of the blue. And I believe in AI as a tool. But this is nonsense.
JoelBarr@maxkennerly I assumed it was just to avoid liability. They can't predict what harm might come of it, so they put the regulators, who can't predict it either, in line to take the fall.