Kevin Foley

PhD candidate in political science. Political speech and ideology in authoritarian politics. Also, ML research. Opinions my own, etc.

[Kevin Foley @KevinFoley]

@BonerWizard If you've only got nine hours watch Secret Agent, Marty Supreme, Bugonia and Train Dreams, make a day of it

Tier two is Sentimental Value and F1, opposite ends of the machines vs emotions spectrum

Frankenstein is a gd snooze

(Haven't seen Hamnet)

[Kevin Foley @KevinFoley]

lol this is not how elections work here at all

[James Rosen-Birch ⚖️🕊️ @provisionalidea]

Democracy is premised on a wisdom of crowds. If everyone votes according to their beliefs and interests, and is honestly represented, governance in the collective best interest results.

It’s not this “plug your nose and vote strategically” thing that’s been around my whole life.

[François Chollet @fchollet]

Interesting work on reviving RNNs. https://t.co/kTOze8qINv -- in general the fact that there are many recent architectures coming from different directions that roughly match Transformers is proof that architectures aren't fundamentally important in the curve-fitting paradigm (aka deep learning)

Curve-fitting is about embedding a dataset on a curve. The critical factor is the dataset, not the specific hard-coded bells and whistles that constrain the curve's shape. As long as your curve is sufficiently expressive all architectures will converge to the same performance in the large-data regime.

[Kevin Foley @KevinFoley]

Hey, here's an ML system that reduces bias

https://t.co/FqKstvgQob

[Kevin Foley @KevinFoley]

@prchovanec It gets harder every day on this app to tell the difference between AI-generated reply accounts, anti-semitic weirdos and people going through genuine mental health crises

[Kevin Foley @KevinFoley]

One of the terrible things about this is you just know some right wing provocateur is going to kill a few cats just to gin up fake evidence. Hard to imagine a scenario where that doesn't happen

[Kevin Foley @KevinFoley]

Vance calling working moms "neoliberal" really a twist here

[Headquarters @HQNewsNow]

Unearthed audio: JD Vance says he agrees that “the whole purpose of the postmenopausal female” is to help raise grandchildren and that helping raise children is a “weird, unadvertised feature of marrying an Indian woman”

[Kevin Foley @KevinFoley]

The potential dangers identified in this bill (critical infrastructure attacks, computers doing human crimes, creating novel biological weapons) are going to remain very do-able for motivated state and non-state actors, regardless of the penalties you impose on AI researchers.

[Kevin Foley @KevinFoley]

A few comments on this debate:

1) The legislation does a terrible job of distinguishing between models and systems. It defines an "AI model" as "a system that can make predictions" but the rest of the text is mainly concerned with restrictions on training of models. Not good!

[Gary Marcus @GaryMarcus]

Dear Fei-Fei [@drFeiFei],

The stakes are really high here so please forgive this open letter.

We've known each other for a long time and I very much respect your work. But your Fortune letter opposing SB-1047 seems off the mark to me, in part because it doesn't fit my understanding of what SB-1047 actually calls for, and in part because it does too little by way of offering a legitimate alternative.

Here are some of my concerns:

• You claim that "SB-1047 will unduly punish developers and stifle innovation. In the event of misuse of an AI model, SB-1047 holds liable the party responsible and the original developer of that model" and in this connection that "It is impossible for each AI developer—particularly budding coders and entrepreneurs—to predict every possible use of their model." But SB-1047 does not require predicting every use.

Rather, it focuses on specific, serious "critical harms" such as mass casualties, weapons of mass destruction, large-scale cyberattacks, and AI models autonomously committing serious felonies. Those seem reasonable to me, and I don't understand what would justify an exemption there. Even then developers are required only to implement "reasonable safeguards" against these severe risks--not to fully mitigate them. Furthermore, much of what would be required is already something companies committed to voluntarily, in discussions at the White House and in Seoul. None of this is really conveyed in your Fortune essay.

• You argue that SB-1047 is potentially "stifling innovation" on the assertion that the bill could harm open-source AI development because of "kill switch" requirements. But as I understand the latest version of the bill, the "kill switch" requirement doesn't apply to open-source models once they are out of the original developer’s control.

• You claim that the bill will hurt academia and"little tech" and put others at a disadvantage to tech giants. But you don't make entirely clear that much of the bill's requirements are limited to models with training runs of $100 million+. Companies that can afford that, presumably valued in the billions, are not exactly "little-tech”, and should be able to handle what is required.

• You say that you favor AI governance, but don't make any positive, concrete suggestion for how to address risks such as mass casualties, weapons of mass destruction, large-scale cyberattacks, and AI models autonomously committing serious felonies. With no other serious proposal on offer, I personally favor SB-1047, though I would welcome discussion of alternatives.

Lastly, asking for standards (and a degree of care) is not unique to AI; it's common across many industries to ask that companies evaluate the safety of their products according to set standards: just look at the pharmaceutical industry, or aviation, automobiles, etc. As Bengio, Russell, Hinton, and Lessig observed, "There are fewer regulations on AI systems that could pose catastrophic risks than on sandwich shops or hairdressers." Your letter doesn’t really grapple with this.

I'm sure that your argument against SB-1047 was made in good faith, and with the best of intentions. But as noted above there seem to be some inaccuracies in your essay, and I wonder if you would be willing to reconsider in light of these clarifications.

Best regards,

Gary
Professor Emeritus, New York University
Founder and CEO, Geometric Intelligence (acquired by Uber)
Author, Taming Silicon Valley

[Kevin Foley @KevinFoley]

That's a nightmare scenario! When I think about the serious social harms of AI, 9 times out of 10 it's labor-related, or it's companies and scammers flooding consumers with spam, bad customer service, crappy advertising, and SEO optimization hell. And ...

[Victor Von Doomscroll @redderbeanpaste]

I want to talk about why Palestine and the American complicity in the genocide there has been such a pivotal issue for me , a China research and policy specialist in dc.

Fundamentally it is about being consistent and conscientious about human rights

[Kevin Foley @KevinFoley]

The alternative is just the objective reality that the Democratic Party is more popular, right wing media is most known for saying and doing bad things, and so on. Constructively, it would be nice to see an extension of this analysis over Wikipedia edit histories (2/3)

[Kevin Foley @KevinFoley]

This is a cool project with some major issues, first of which is the framing of the finding as inherently negative or "biased." Should we be so surprised that the party that has won nearly every national popular vote in my lifetime is more associated with positive sentiment? (1/)

[David Rozado @DavidRozado]

1/ Is Wikipedia politically biased? To explore this question, I averaged the sentiment (negative/neutral/positive) associated with a set of politically loaded terms used in English Wikipedia content (N=175,205 sentiment annotations).
https://t.co/8c14FISRos