Comment by prionassembly | Hacker News Reader

Comment by prionassembly | Hacker News Reader

Parent Full thread

Are we evaluating her as a researcher or as an activist?

I’ve read some of her research (that’s on arxiv) and it’s... not on the level that “Google AI researcher” implies. It’s good that they hire from the Humanities, but we tend to smush all academics as if they all had equal rigor and value.

Maybe she’s a great activist. Or a great cook. Being it that I’m not into activism, I can’t judge that.

----

Edit: I have to hijack my sudden spike in karma (and, I assume, visibility) to "amplify" some important research milestones in the Humanities that impact the problem of "AI ethics" taken seriously.

[General]

* Baudrillard - "Forget Foucault"

* Delanda - "Philosophy and simulation"

[Science and technology issues]

* Latour - "The Pasteurization of France"

* Latour - "Reassembling the social"

* Caillon - "Acting in an uncertain world"

[Computers and society]

* Various authors - "Software Studies: a lexicon", available here https://monoskop.org/images/a/a1/Fuller_Matthew_ed_Software_...

>It’s good that they hire from the Humanities...

Gebru has a BS and MS in electrical engineering, and a PhD in computer science, specializing in artificial intelligence; all from Stanford. After receiving her MS, she worked at Apple's hardware engineering group. After her PhD, she worked Microsoft's AI research lab.

That's... surprising. I've read (really read) a handful of her papers on arxiv, and the only that isn't a gooey Problematic Studies fondue has ten coauthors (with her in 5th or 6th place). She really has the lingo down pat.

Maybe the problem is precisely that they didn't actually hire from the Humanities? Although the cynic in me thinks they just hired someone with a "SJW" discourse rather than find themselves criticized for their market strategy -- whether their use of their monopoly power is ethical.

As it stands it really sounds like they got someone to offer Yann LeCun as a scapegoat for AI products (Amazon Rekognition is infamous); these services then serve as scapegoats for police departments, which...

> Are we evaluating her as a researcher or as an activist?

I honestly don't know how to take this comment.

Are you comparing her work to cultural theorists like Baudrillard, etc.? Because a brief read through her papers and they seem to be dealing with common-sense ethical concepts, not French social critique.

Otherwise, this comment is just odd - no real specific criticism of her body of work except to suggest that it is less good than what you'd expect from a "Google AI researcher".

I'm suggesting that HN readers who are upvoting my comments familiarize themselves with these authors.

Gebru is a successful "scholar" as far as the credentials game goes; I don't presume to have anything to teach her.

Edit: Are we sure calling something "common-sense ethics" is a common-sense proposition? Are we comfortable deriving this Ought from that Is?

> Gebru is a successful "scholar" as far as the credentials game goes; I don't presume to have anything to teach her.

Again: back up your assertions or they are worthless. I'm reading her papers and have also read many of the authors you are mentioning, there is very little in common.

> Are we sure calling something "common-sense ethics" is a common-sense proposition? Are we comfortable deriving this Ought from that Is?

So your critique now is that her papers are steeped in a moral realist perspective? It seemed you were going for the dogwhistle-y "diversity hire" rhetoric, but I'd welcome any actual critique on the merits.

You're starting to sound like you're writing in good faith.

I'm not proposing a critique of her work. What I've found on arXiv reads like standard fare for its genre. Hotelling famously joined an early seminar on linear programming with Dantzig only to interject "... but the world is nonlinear!" That's not what I'm out to do -- not with this particular researcher, nor with its proxies on twitter and HN.

The scenario I'm looking at is one where the polarities are not left versus right or woke versus straight, but Humanities versus STEM. If people are to reject her kind of work -- and I'll be glad if they do, even if I'm not really willing to press them in that rejection with a proper "critique" -- then they could walk away burned by theory and postmodernism and all such crap. I'm proposing an entirely different path, or a space of different paths even.

If you're referencing Baudrillard (points to HN profile), Simulacra and Simulations ought to go up there too, for this technology crowd.

Reply on news.ycombinator.com