Live data from Hacker News

Microsoft lays off one of its responsible AI teams

platformer.news

51–60 of 466 posts

Re: Microsoft lays off one of its responsible AI teams

#51

These sorts of adjunct, off-to-the-side "ethics" teams are always corporate window dressing, and when push comes to shove they always get jettisoned when competition demands an increase in speed. I don't mean to start a flame war, but this is a reason why I think "Chief Diversity Officers" are such an ill-conceived notion. It's not that I don't think diversity in corporate environments is extremely valuable, but it's…

A Chief Diversity Officer owns a slice of the HR organization dedicated specifically to DEI issues, the same as a CTO in a (non-tech-industry) firm owns a specialized slice of the information organization.

Re: Microsoft lays off one of its responsible AI teams

#52
post #34

These sorts of adjunct, off-to-the-side "ethics" teams are always corporate window dressing, and when push comes to shove they always get jettisoned when competition demands an increase in speed. I don't mean to start a flame war, but this is a reason why I think "Chief Diversity Officers" are such an ill-conceived notion. It's not that I don't think diversity in corporate environments is extremely valuable, but it's…

What about Chief Compliance Officer or Chief Risk Officer? (Granted, not an awesome moment to be arguing that those are effective...) Or Chief Security Officer?

Excellent point actually, I think those examples made me adjust my opinion a bit. Chief Security Officer I don't really agree is analogous, because while all teams are responsible for security, there is enough "independent security work" (e.g. managing infrastructure and IT security) to warrant their own teams.

But I think Chief Compliance Officer and Chief Risk Officer are pretty analogous - their main directive is to ensure other teams adjust their behaviors to follow the rules. I think the difference, though, is that these roles have unambiguous, and in many cases legally defined, requirements. I actually think it is the right moment to argue those are effective - SVB famously went without a Chief Risk Officer for a year despite having an unusually high number of risk committee meetings. We've all seen what happens when companies don't take risk and compliance seriously.

With things that are "fuzzier" (and by "fuzzier" I mean real but ill-defined/hard to measure impacts on the business of making money) it's just too difficult to have impact on teams you don't own. A Chief Risk Officer can easily make the case "If you don't do what I say, our company will fail - like, completely cease to exist in 2 days." I don't think a Chief Diversity Officer has the luxury of that sense of immediacy.

In any case, thanks for helping me think about this a little more broadly.

Re: Microsoft lays off one of its responsible AI teams

#53
There’s a danger of confusing the problems with actually existing AI ethics work for evidence that there’s no need to ethically evaluate AI.

On any moral realist view, AI will likely be extremely important. And insofar as ethical standards are used to organise society, ethical views on the use of AI will likely be developed in the same way as they are at least very weakly developed and observed in most other fields, so an understanding of them will be important for understanding society simpliciter.

Re: Microsoft lays off one of its responsible AI teams

#54
post #17

Earlier quoted context omitted.

"If I don't make the world eating machine, someone else might make the world eating machine first!" No wonder we don't see aliens filling the universe. Intelligence is its own great filter.

How can you stop a computer program? Make math illegal?

I mean if you want an honest answer, nuke any place that's making GPUs/TPUs. I mean, they don't spring up out of the ground and at least our current models require insane amounts of GPU/TPU based compute. The factories are fragile billion dollar facilities, well beyond the difficulty of extracting uranium.

Now, that probably just delays the world eating machine for some time, but it would be a hell of a lot harder to make anything like our top of the line models without access to massive amounts of both compute and data.

The hard problem isn't making the world eating machine. The hard problem that silicon valley doesn't give a shit about is making the world eating machine even safe enough to be on the same planet as without it melting the entire crust when you say "oh, it's a little chilly out today".

Inner alignment is completely unsolved and the pace at which we are solving any alignment issues is much, much slower than the rate we are making gains in intelligent unaligned machines.

Re: Microsoft lays off one of its responsible AI teams

#56
post #30

Earlier quoted context omitted.

[flagged]

I do think the push for ethics in AI is important. As much as these functions do get sidelined in the prioritization process as outlined in the parent comment, the intent of introducing the rigor is valid.

Thing is, what gets called "AI ethics" actually has approximately* nothing to do with AI.

It's all about whichever field the particular use-case belongs to, and would lose nothing by seeing the AI as a black box.

* IIRC there have been issues with models that find correlations being presented as finding causality. That is something that belongs to AI as a field; other people can't evaluate your stuff accurately if you've misinformed them about its capabilities.

Re: Microsoft lays off one of its responsible AI teams

#57
post #17

If Microsoft and Google don’t release, some startup will beat them to it. Open source means raw AI will become widespread. It’s just math anyway, and you can’t stop progress.

"If I don't make the world eating machine, someone else might make the world eating machine first!" No wonder we don't see aliens filling the universe. Intelligence is its own great filter.

Not in this case. If intelligence always created intelligent machines that are capable of replacing us, then how is that not intelligence as well?

Further, silicon-based lifeforms would be able to spread and travel the universe much easier because what is time to a machine?

Re: Microsoft lays off one of its responsible AI teams

#58

Earlier quoted context omitted.

Judging from what happened when they fired all their qa people…

…They released Win10 about a year later. My Win10 experience was pretty good overall, from as early as I updated to it. (I wasn’t the earliest adopter, but I was using it sometime in 2016.)

From the biannual major updates to the monthly Patch Tuesdays, Windows 10 has been plagued by more or less glitches. This has been attributed to Nadella’s decision to abolish QA. This problem persists in Windows 11, which caused issues with mouse cursor movement in yesterday’s cumulative update.

Re: Microsoft lays off one of its responsible AI teams

#59
post #27

All the AI ethics stuff seems divorced from reality anyways. Waxing poetic doesn't actually fix the world.

It is divorced from reality. There's only one north star in a capitalist mode of production, and it's the profit mechanism. Corporate ethics is a pipe dream. The only reason to have such a team is because their existence would shape perception enough to create profit either through motivating the workers ("we can't be doing harm, we have an ethics team backing us") or through PR ("Microsoft announces ethics specialists on new AI offering to answer any concerns over moral issues blah blah").

Re: Microsoft lays off one of its responsible AI teams

#60

These sorts of adjunct, off-to-the-side "ethics" teams are always corporate window dressing, and when push comes to shove they always get jettisoned when competition demands an increase in speed. I don't mean to start a flame war, but this is a reason why I think "Chief Diversity Officers" are such an ill-conceived notion. It's not that I don't think diversity in corporate environments is extremely valuable, but it's…

But isn't the idea of a semiautonomous person/group overseeing ethical concerns precisely to overcome the very problem you're pointing to, namely, to not make their decisions based on a competitive edge? In an alternative top-down organization like you are saying, why would the rational company leaders ever pick the ethical side of a dilemma if the other lemma is making more money? Or hire anybody who would make that "wrong" decision?

Like I can't understand this mindset where one says: "its clear the ethics team is getting in the way of us doing the business we could be doing, so its clear they are the problem." Isn't the friction the ethics team produces precisely their raison d'etre?

Like at what point can we all stop pretending that we are angry at woo-woo philosophy experts telling us "no," and admit we really just have a conscious which is at odds with overwhelmingly financial goals of this world.

Post reply on HN