Why I Left Google DeepMind

Posted by eatitraw 7 hours ago

Counter144Comment37OpenOriginal

Comments

Comment by vintagedave 6 hours ago

Very long to read, but highly principled. I feel his desperation. Powerful people can too easily do little, and ignore their consciences. I appreciate he named those he spoke to; it came across as fair, and makes them accountable for what they did or did not do.

Comment by fuzzfactor 2 hours ago

>Powerful people can too easily do little, and ignore their consciences.

IOW:

>Congress remains a potted plant in the corner.

Comment by mark_l_watson 2 hours ago

Great show of principles and character, +1000 to the author.

What can we "regular people" do?

My approach, which is very inconvenient, is to try as much as possible to not feed my consumer spending money to tech giants. In some cases I fail at this (e.g., my wife and I will always buy a family membership for YouTube Plus with music; we still order some things online from Amazon) but I make an effort.

I have been an AI practitioner for over 40 years and I want to use the very best AI coding agents, etc., but I am forcing myself to just use access to cheaper open weight models from smaller US inferencing companies (my last $$ subscription runs out in a few days).

Comment by hmokiguess 5 hours ago

However way people that are deep into the cog wake up I support it. Once the truth was seen you didn't bend and stood by your principles, it must have been very challenging to navigate that from the inside with your entire identity tied to and associated to the thing you ended up growing out from. Appreciate you, welcome to your new life. May you find peace and continue to do good work next.

Comment by tdb7893 5 hours ago

I'll always support someone taking a personal stand about what they can or cannot do. Ethical issues always invite the most criticism but my experience is that it's rare for anyone to really give anything up for their ethics (this person likely has money saved up and will land on their feet but even this situation is more than most people would do).

Comment by aleph_minus_one 5 hours ago

Primary source: https://turntrout.com/why-i-left-google-deepmind

Comment by adverbly 4 hours ago

If you want to do good in the world: You are not alone!

Its not easy to do at this moment in time, but be patient, consistent, and spread your attitude to others.

Comment by ChrisArchitect 3 hours ago

ICYMI

Previously on the author's source: https://news.ycombinator.com/item?id=48925271

Comment by 1-6 5 hours ago

The virtue signaling is too much on this article. It could have been a well-written piece, but they had to make it about moral grandstanding.

Comment by layer8 3 hours ago

“Virtue signaling” is when you don’t take meaningful action. The article author went out of their way to take meaningful action.

Comment by woopwoop 3 hours ago

I'm like the Japanese soldier who kept fighting into the 1970s on this, but actually signalling is about taking costly action. When people call something virtue signalling what they usually mean is that it's cheap talk, but this is wrong. The whole point of signalling is that it's costly. A more correct application of virtue signalling would be something like those people who contort their lives to genuinely try to produce zero garbage. This is a truly costly action, but it is signalling in the sense that it is more about showing that one cares by being effortful than actually doing something effectual.

Comment by dpark 1 hour ago

> When people call something virtue signalling what they usually mean is that it's cheap talk, but this is wrong. The whole point of signalling is that it's costly.

What? Signaling is not costly. It’s literally a display, a signal.

Comment by vovavili 18 minutes ago

Going out of your way to quit a job most people in tech dream of after months of serious calls to action is not mere signalling.

Comment by vintagedave 4 hours ago

Being ethical is the point.

Comment by golly_ned 4 hours ago

Any case where someone takes a moral position is now called "virtue signaling" and "grandstanding."

Comment by leoh 19 minutes ago

I am not sure how to square the posts that AI is a bubble with the ones that say AI is so terrible; lately I am somewhat inclined to especially believe the former.

Comment by fragmede 5 hours ago

more comments at https://news.ycombinator.com/item?id=48925271

Comment by readthenotes1 6 hours ago

I am outraged to find gambling in this casino!

What sort of ethics bubble was the author living in to believe that Google was anything but solely focused on achieving shareholder value?

Comment by zbentley 2 hours ago

I don't know the author, but I do know that it's very much possible to work for organizations whose values and ethics are extremely compromised, and to mindfully participate in their work because you can also do a lot of good for a lot of people while you're there.

It's a complicated cost/benefit tradeoff: are the opportunities for world-improvement available with the resources of an ethically compromised organization worth the world-damage done by your participation (either your presence alone lending legitimacy/size to the org, or by work you may be called upon to do that directly or indirectly does harm)?

If you think there is a simple answer to that, you're wrong.

Like, sure, a lot of people delude themselves into thinking they're not doing harm because they like money or hate change. That's bad, and separate from what I'm describing above: that you can work for a deeply flawed org and accept the cost of your participation in those flaws in return for benefits that you can others can achieve with the resources offered by that same flawed org.

Comment by tmtvl 5 hours ago

Agreed: expecting Alphabet not to do the evil greedy thing is like expecting a lion to become a vegetarian, it just isn't going to happen.

Comment by Ygg2 5 hours ago

I wonder if in a few years, there'll be heads of Google paraphrasing the old skit "You know, in hindsight, allowing autonomous AI drones to carry fusion bombs was a bad move, but for that one quarter we really earned a lot of shareholder value"

Comment by pjmlp 5 hours ago

The same that believes in HR marketing like Do No Evil.

Comment by Ygg2 5 hours ago

Pretty sure they ditched that motto long time ago.

Comment by andriy_koval 5 hours ago

Still there:

And remember... don’t be evil, and if you see something that you think isn’t right – speak up!

https://abc.xyz/investor/board-and-governance/google-code-of...

Comment by pjmlp 5 hours ago

They did, still doesn't change the kind of bubble we are talking about.

Comment by skew-aberration 5 hours ago

Not only that but his outrage is narrowly focused on contemporary US politics. Billions of people access food, shelter, medicine, education, etc, only by the rapidly diminishing leverage their labor provides (or doesn't provide) them. These people presumably never so much as picked up a book, or evenly briefly entertained a thought about a person outside their own immediate social circle.

Comment by iluvcommunism 5 hours ago

[dead]

Comment by meta_gunslinger 6 hours ago

[flagged]

Comment by nmatej1 7 hours ago

[flagged]

Comment by gyanchawdhary 5 hours ago

painfully entitled and arrogant

Comment by golly_ned 4 hours ago

For having a moral position on a moral topic?

If he were strictly self-interested we'd have no objection to it in this culture.

Comment by cloudie78 5 hours ago

Wait, only now you realised that the “do no evil” meant jack shit?

How blind and ignorant really are you?

Let me attempt to save you from another horrifying realisation: it has hardly anything do with who’s in the office right now.

The horrors and evil will persist.

Comment by fuadnafiz98 6 hours ago

Someone send this to Tucker Carlson

Comment by negamax 6 hours ago

You (the org) brought fire to the masses. How that fire is used (to build or burn the cities) is not for you do decide

Stand by and share your ethics concern. But this isn't misalignment. A lot of AI use cases will be constraint/reigned in through governance evolution at every scale

Comment by tw04 5 hours ago

Comparing it to fire isn’t really an apt comparison.

You developed the nuclear bomb. Would you give free access to every human on the planet to do with it as they see fit? Or would you demand responsible guardrails from the start?

AI companies want it both ways. This will both redefine humanity as we know it, but also nobody should try to enforce any guardrails. Why? Because it might hurt profits.

Either it’s snake oil and not as revolutionary as claimed, or people should be facing prison time for trying to claim “I just invented it and gave everyone access, I can’t be held responsible for what happens next”.

Comment by leoqa 1 hour ago

I mean, does enabling ICE to be more effective actually represent a harm? ICE has authority to do the things it’s doing. The courts are the appropriate venue to check that authority. Two agents murdering citizens is a criminal matter completely unrelated to Google’s relationship with DHS.

It feels like AI empowering scams and facilitating post-truth is more harm.

Comment by negamax 5 hours ago

Nuclear wasn't governed until the destruction was seen. There's no reason to believe AI wouldn't follow a similar path

Comment by tw04 5 hours ago

The bomb was literally developed under a top secret program, and they didn’t hand out functioning bombs to the general public at any point in time before “the destruction was seen”.

Comment by negamax 5 hours ago

It was created at wartime. Not as a civilian project. Developed as a weapon. And other countries were developing nuclear weapons as science of it was never a secret

Destruction display did caused the governments to come together and build frameworks and structures to avoid the repeat of it

Comment by tw04 4 hours ago

Which again, is completely different than giving access to every human being on the planet, then trying to figure out how to regulate it later.

Out of the gate access was restricted to government use, and a limited subset of governments at that. There was never a point in time it was handed out to any private citizen to do with as they pleased.

As for “we haven’t seen the destruction yet” - huh? Between the proliferation of hacks utilizing AI and drone warfare in ukraine we’ve ABSOLUTELY seen the destruction already.