Timnit Gebru (she/her).

When this "AI" bubble pops, the men pretending they weren't pushing the hype, like "critics" whose position is "AGI is real but LLMs aren't the way," who were in eugenicist and "AI existential risk"🙄 circles, will get specials discussing what they saw coming, when its the women who actually told you so.

And then it will be rinse and repeat, onto the next grift.

Dec 10 at 8:53pmWeb
Show threadalcinnzDec 10

@timnitGebru You know: I wonder what the next grift will be! So I can be prepared to refute it.

It looked like it was going to be quantum computing, but that seems to have fizzled out.

I somewhat doubt you have the answer...

Show threadTimnit Gebru (she/her).Dec 10
@alcinnz No idea. I only know the current grift but I can't quite cash in to follow through on it 😂
Show threadIsaac Ji KuoDec 11

@alcinnz @timnitGebru Well, the "AI" grift has ruined the term "AI" (which applies to way more than just LLM and stable diffusion)...

Maybe the next grift will ruin the term "AGI"? (Not that there's as much to ruin there.) I dunno ... something named "Basil" as in Roko's Basilisk except Basil just outputs random phrases via SQL query so it's actually efficient.

The point is, you must integrate BasilAGI into everything or else future Basil will punish a simulation of you.

Show threadShadowInTheVoidDec 11
@isaackuo @alcinnz @timnitGebru this is actually about the 4th "AI" boom, and each time "AI" has ment something different that was being oversold. This one has just been bigger and louder. Let it burst and watch "AI" be a dirty word for a while and then in 10 years they'll be another thing being labled as "AI" again
Show threadOldGeekDec 11
@alcinnz @timnitGebru I think the Republican party tying the $ to crypto. Just waiting for the trillion dollar rug pull.
Show threadDianeDec 16

@alcinnz @timnitGebru

I bet the next grift will be the longest running grift, selling "the reason you're suffering is because of that group over there that you've always hated and not because you were scammed"

Show threadFabian TranschelDec 10

@timnitGebru You know, I get your point. It's as valid as it's painful.

But please don't phrase as ALL men vs. ALL women. There *are* men that called BS just like there *are* women that fueled the hype.

Show threadDavid CulleyDec 10
@ftranschel Where did she write what you tell her not to write?
Show threadDaniel Terhorst-NorthDec 10

@ftranschel with respect, please don’t not-all-men @timnitGebru, of all people.

To a first approximation she is absolutely correct.

Show threadfromjason.xyz ❤️ 💻 ✍️ 🥐 🇵🇷Dec 11
@ftranschel @timnitGebru Idk man, I felt zero pain reading this nor did I think that I was personally included in the statement so maybe this is just a you thing? Take a lap. Get back to us
Show threadTedDec 11

@ftranschel I get a little defensive about phrasing. But I don't think there's any remote implication here that "all men", even "many men" are the problem. This reads to me very much as "certain men"

There isn't a claim that there aren't women taking the same problematic stances. This is about who will be awarded the spotlight, even though they contributed to the hype. And it will probably be men. Just because that's who historically are given the most voice.

Show threadPalmAndNeedleDec 10
@timnitGebru block chain, Uber for XYZ, pivot to video, ... The list is long
Show threadEm Dec 10
@timnitGebru @dalias 
Show threadmccDec 10

@timnitGebru I'm increasingly worried the annoying forced AI features won't even actually get removed from the operating systems/websites we have to use. Maybe stuff will just get frozen in place, no rehiring, no fixing the stuff that broke during the AI era.

(I don't understand how anyone but Apple *could* continue supporting genAI features once "tokens" are no longer subsidized but the market seems to be capable of any level of irrationality.)

Show threadfromjason.xyz ❤️ 💻 ✍️ 🥐 🇵🇷Dec 11
@mcc @timnitGebru I agree. I think they're keep shoving it down our throats long after the bubble bursts. They'll just make the features worse, and bundle it with services we actually need with a 25% hike in price. If they get their AI license laws and local models get outlawed then they'll really be cooking.
Show threadDavidDec 10

@timnitGebru

Next grift, is train a LLM with Trump speeches. So they can pretend his consciousness was uploaded into the cloud, making him inmortal, and rule America for all eternity. 😏

Show threadjandiDec 10

@DBG3D @timnitGebru Have you seen this? https://www.lawfaremedia.org/article/the-unitary-artificial-executive
It made me put a different face, more like 🥶
Granted, it's from someone enamored of "AI", almost giddy, but still, chilling as fuck.

Via Executive Functions Chat: Executive Power in the Age of Artificial Intelligence

Episode page: https://www.execfunctions.org/p/executive-power-in-the-age-of-artificial

Media file: https://api.substack.com/feed/podcast/181137725/f7b1bb02ccb18956832b58c64c394151.mp3

The Unitary Artificial Executive

Previous expansions of presidential power were still constrained by human limitations. Artificial intelligence eliminates those constraints.

Default
Show threadDavidDec 11

@jandi @timnitGebru

And that is why this con with LLMs and AI must be stopped. 🤷🏽‍♂️

Show threadMarius Gundersen - mdg 🌻Dec 11
@DBG3D @timnitGebru more than a decade ago I made a markov chain bot trained on all the fluffy corporate emails from our CEO, which we had great fun with generating stupid text in Slack. My point is, this is possible today, and would convince the people needing to convincing, which makes this prediction very scary.
Show threadDavidDec 11

@gundersen @timnitGebru

I'm sure I'm not the first one on having such idea. Remember Nixon's head in futurama? 😏

Show threadI can’t believe its not @twcauDec 10
@timnitGebru And a lot of good people will hopefully have kept the receipts, so that cohort can’t escape their stupid.
Show threadsammypandaDec 11
@timnitGebru trueeeee
Show threadyeahyeahyensDec 11
@timnitGebru and -surprise - guess who will have to clean up after this mess
Show threadVolineDec 11

@timnitGebru

Bubbles and grifts are all capitalism has now.

Show threadCaptain ObservantDec 11
@timnitGebru All I want for Christmas is an AI bubble bursting, 7 stocks tanking, 6 companies bankrupting, Bitcoins collapsing, etc
Show threadJDec 11

@timnitGebru
Aren't you mixing things?

AGI will be real some day but LLM are not the way (LeCunn), hence the current trend IS indeed a bubble.

Show threadskaphleDec 11
@darkphysics @timnitGebru What makes you say "AGI will be real" instead of "AGI might be real"?
Show threadJDec 11

@skaphle @timnitGebru

Just quoting LeCunn in X.

My personal view is the same, current LLM are lacking many human skills.

Show threadskaphleDec 11
@darkphysics @timnitGebru Ah okay. I wanted to ask whether there was a deeper argument behind that. Obviously that person is who Timnit addressed, so she sees it different. Two different and opposing arguments need to be checked outside and you can't use one side to say the other is wrong. Does that make sense? Maybe, if you follow what this LeCunn person says, could you rephrase the argument they make? Or what brought you to that same view?
Show threadJDec 11

@skaphle @timnitGebru

Good point on the argumentation but LeCunn is the creator of CNNs, and helped in many other AI architectures, that's why I wanted Timnit to give a more detailed explanation.

https://en.wikipedia.org/wiki/Yann_LeCun

What is his (deeper) view on the subject? I can only guess:
LLM at their current state are a sequence to sequence generator, no evolving nature, no "new" knowledge, no ability to set own new goals, forgetful, no continuous learning, etc, etc

As I said, just guessing.

Yann LeCun - Wikipedia

Show threadskaphleDec 11

@darkphysics @timnitGebru Thanks for the clarification. Your question had appeared to me more like a question of her conclusion than a request for explanation.

I understand the limits of LLMs and that probably there's no AGI coming out of it. And I think that is the consensus here. As I understand it, the diverging views are over whether other methods, like CNNs, can lead to AGI. My position is I don't know, but I'd like to know the reasoning for one or the other sure statement. It seems to me Timnit is sure that it's so questionable that people who argue for it are lying grifters. And it seems to me that you, or LeCun, are so sure that you make statements that there *will* be AGI.

If you don't want to elaborate, that's perfectly fine too.

Show threadHiisikoloartDec 11

@timnitGebru
It would be fun to see my ex-friend possibly crawl back once the bubble bursts and try to apologize for abandoning our friendship to defend "AI" theft. I would really enjoy seeing him not only realize what he did wrong, regret it, and need to live with the consequences of it.

(I am an artist, and I suspect he was feeding my art to an LLM and reacted horribly because of an unrecognized guilt for it, and for replacing me in a project with "AI".)

Show threadTimo GrünDec 11
@timnitGebru
It's so obvious that all current models are high level pattern matching and that there is zero cognition or reasoning going on. At the same time Sam Altman and the other LLM bros continue to bloviate about AGI, that they cannot possibly achieve on this path, with most of the media falling in behind them like congregations wowed by the miracle mechanisms of the medieval and ancient priesthoods. With the ludicrous sums being invested there's no way this can end well.
Show threadFrank Bennett 🇯🇵Dec 11
@timnitGebru Is there a next grift though? Seems like pretty slim pickins out there.
Show threadPalmAndNeedleDec 11
@fgbjr @timnitGebru there is always a next grift
Show threadMathieu PeronaDec 11
@timnitGebru I strongly relate. Some people are still riding a "I predicted the 2008 subprime crash" reputation, although the imminence of such a crash was announced years before by a confidential news magazine. You know, The Economist (front page, at that).
Show threadFlipper 🐬🏳️‍🌈Dec 11

@timnitGebru I read a cool article in late 2028 that said Yann LeCun first coined the term "Stochastic Parrot" in 2022 -- so prescient.

It's why I'm stoked he's starting his new Soylent company, bro.

Show threadAdam KatzDec 11
@timnitGebru we can look to cryptocurrencies to see where this is going: the grifters keep doubling down, embedding their crap everywhere, building a dangerous house of cards that will take down other industries.
Show threadRainbow StardiverDec 11

@timnitGebru it won’t just be them. The big media outlets will also prefer to elevate people (especially men) who profited along the way and claim they finally see the harm, instead of people (especially women) who were critical the whole time. Maria Farrell’s Prodigal Techbro piece from 2020 is relevant here.

https://conversationalist.org/2020/03/05/the-prodigal-techbro/

Credit: @mariafarrell

The Prodigal Techbro

Prodigal tech bro stories skip straight from the past, when they were part of something that—surprise!—turned out to be bad, to the present, where they are now a moral authority on how to do good…

The Conversationalist
Show threadErik JonkerDec 11
@timnitGebru ...the AI bubble will pop, but it will primarily be a financial crisis, GenAI will remain and be used everywhere. Although maybe more responsibly etc
Show threadJackDec 11
@ErikJonker what makes you think that the technology will stay prominent? The result of over-speculation may not take down the operations of the industry, no, but I do not see a way for OpenAI and others to remain profitable in their current state. The outputs of current LLMs are just not helpful when you consider the resources needed to produce those outputs, and companies have repeatedly stated they are not willing to pay for them.
Show threadErik JonkerDec 11
@technotion ....many medium-sized LLMs are not that powerhungry, are already being implemented, also in on-premis/local applications. Also Opensource models are gaining momentum, surely things will change after a financial collapse but the technology and it's usefull applications (yes, they exist) are here to stay.