Nothing but respect to TurnTrout for taking an action like this. The world needs more smart people who are willing to stand for what they feel is right, despite the pressures otherwise. Without that occurring more, our species is going to lose many impactful prisoner's dilemmas coming these next two decades.

This raises my respect for AI researchers a little bit too. I have often felt that the entire industry is pretty tainted to the core, and for better or worse that colors my opinion of the researchers.

Maybe I'm in the minority, but I thought it was gross to download pirated art for a student project when I was at Berkeley years ago. So it has been really sad to witness many of the most brilliant minds of this generation answering the siren song of disrespecting the collective effort of others to extract and resell residual value.

I'd guess TurnTrout doesn't agree on that framing, otherwise he probably would not have been at Deep Mind. But clearly he and I agree on other ethical positions; I am nothing but glad to see him stick to his principles here.

Props to the author. I left Microsoft due to their work with Israel to spy on Palestinians and record all of their phone conversations (in addition to other IDF collaborations). They ended up walking some of it back, but Satya was complicit.

https://www.theguardian.com/world/2025/sep/25/microsoft-bloc...

Respect for sticking to your beliefs. Just out of curiosity though why do people not want smart AI weapons? I would much rather have an onboard AI that can discriminate between unarmed civilians and military assets, seems irresponsible to not... Is a dumb sea mine that blows up everything somehow better than a smart sea mine that knows to not blow up sometimes?

I haven't finished reading, but I should note for the group that the author doesn't represent the Anthropic situation accurately. Emil Michael, the undersecretary of war/defense for research and engineering went on the All-In podcast (1) and explained, for quite some time, exactly what happened. You should go listen to it (2). In essence, in the negotiations, Dario kept coming back saying "well, if you need it, call us, we can redline things as needed". This happened over and over and over. Emil's point was that a major conflict, if it were to occur, might happen on the 30 minute clock of an ICBM, and all due respect to Dario, the national security apparatus just doesn't have an allowance in that 30 minutes to call him for a redline. Emil felt Dario had demonstrated he wanted ultimate control via line item veto, and was willing to trade up to and including the survival of the nation for that veto. And a government cannot be expected to pay for that sort of behavior from an entity domiciled in their jurisdiction.

Now, Dario is going to win a decent slice of the economic pie. But as an military acquisitions matter, I gotta say, I have to agree with the undersecretary's position here, and yeah, it makes sense to document a company's undesirable behavior, and in certain circumstances push that information to others. Not the first time the government has found a company under contract acting in a way that appears to be counterproductive to the government's obligations; there's a whole database full of this stuff (3).

1) Without a doubt, All In is a friendly crowd for Emil, but I think that actually made it easier for him to get more nuanced facts out in this case because he wasn't spending a lot of time defending malignant attacks.

2) This link should jump to the relevant segment https://www.youtube.com/watch?v=gzwRflcLPAA&t=2479

3) http://www.ppirs.gov/

Principled people have become so rare

I did not know about TurnTrout till now and thanks to HN community now I know. I am beyond impressed as to how he lives by his principles. I don't think I have this kind of courgae or confidence and thankfully I don't work for Google or have to quit it. On the other hand we have smart people who have choose to work for companies like Palantir and are proud of it.

This is inspiring, thank you to the author, it must have a hard piece to write.

If the author were to read these comments, I'd commiserate with them on the disappointment of the utopia left behind in the service of acquiring wealth and power. An abundant solarpunk future enabled and augmented by AI, hyped so well by early Google and Kurzweil, captured as capital to earn a buck for shareholders.

Mad props for pushing as far as they could push.

I think a man who makes that kind of choice is seriously admirable. I could never have made that choice myself.

> Plus, a pledge is only worth the credibility behind it. When someone signs “I will not support the development of lethal autonomous weapons,” then stays while their company sells unrestricted AI to a military that wants exactly that, they teach every counterparty a lesson: these safety people will not act, even at their own brightest line. The next commitment they make is worth less. Eventually it’s worth nothing.

This is why I haven’t taken the internally employed (and visibly public) AI safety people seriously for years: they always back down, because they value that fat paycheck and fiscal rewards and associated status over their proclaimed ethical values, red lines, and boundaries. Their actions betray their every word.

That’s also why I push that we remember these names, the ones who fought back as well as those who lied through their teeth. We should remember everyone willing to harm others to protect themselves, and hold them accountable for their misdeeds. In an ideal world, people like Jeff and Sundar and Stuart would be blackballed so hard they’d never be able to get so much as a help desk job ever again, nevermind any position of leadership. It’s why I proclaim them to be monsters, giddily sacrificing the anonymous other for personal vainglory and the avoidance of standing on any sort of principle.

The OP did everything right and demonstrably proved the entire apparatus is rotten to its very core, to the point of infesting and poisoning the very entities professing to stop its worst excesses (IASEAI). Why we give these ghouls the slightest bit of credibility anymore is beyond me.

We know now that Google is a sycophantic company whose DEI initiatives were all fake and dropped once Trump got elected.

This somewhat naive initiative was bound to fail. The good news is that the AI military products won't work, except perhaps for blowing up a girls' school.

Here are CEOs falling over themselves to support Hassabis' regulatory capture proposal:

https://xcancel.com/sundarpichai/status/2077086951833063580#...

https://xcancel.com/satyanadella/status/2077063479232795024#...

https://xcancel.com/elonmusk/status/2077415601610297535#m

It is an exclusive club and we are not part of it.

> Even if Google had adopted your Framework, the Pentagon would have refused

> I agree. xAI would still have given over their AI. But if Google had given signs of independence earlier, it could perhaps have built a coalition with OpenAI and Anthropic.

This is like saying that Google and Apple might build a coalition to prevent App Store regulation. These are competitors, they all see moral flexibility as an advantage. They're not going to take a moralist stance if their federal protection is predicated on federal cooperation. All of these FAANG businesses have already bent the knee in anticipation of this, xAI is just riding the coattails of the federal quid-pro-quo.

When you go to work at a megacorp, you're always leaving your ethics at the door. Yes, there's an attractive pie-in-the-sky fantasy that Apple does care about human rights, or that Google isn't evil, but they're always just lies. I disagree with a lot of GCP's customers, but I'm still shocked that a DeepMind employee would make it this far in the career pipeline before seeing the soylent green get made.

am I the only one that finds it very hard to read this font?

> I wanted AI ethics commitments to hold under pressure.

LLM chatbots, they way they are trained and have data collected for them today, are fundamentally unethical, regardless of whether Google sells services to the DHS.

No less importantly, Google's is a fundamentally unethical entity. After all, it says so right in its motto: "Do Evil".

Ok, that's not it's motto, it's just a pun on the old motto Google dropped when it became too starkly opposed to its practices, of mass surveillance and manipulation and censorship of search and content discovery results for commercial and political purposes, on its behalf and that of governments. This has been widely reported upon.

I will just give a few links regarding Google's complicity in the ongoing genocide in Gaza, as described by the UN special rapporteur for the Palestinian Occupied Territories:

https://novaramedia.com/2025/07/02/tech-giants-and-british-b...

(or get the report itself: https://www.ohchr.org/sites/default/files/documents/hrbodies...)

If you do not like what Google is doing, just leave as I said before [0]. They don't care about you.

[0] https://news.ycombinator.com/item?id=47931336

> Jeff Dean ... and other senior employees pledged to “neither participate in nor support the development, manufacture, trade, or use of lethal autonomous weapons.” Google signed the classified deal, yet they remain.

> I think he could have stopped the deal, yet he did not. He remains, yet I think he should not.

Is Jeff Dean still considered a saint at Google? If so, how come this doesn't change that? The amicus was enough?

tldr: scientist discovers reality

I mean, on the one hand, I support this. But this is going pretty far. When it comes to non-US police and military, if your threshold is killing 2 innocents (and the officers getting away with it), then you can't do business almost anywhere ...

With essentially any country, and even vaguely similar circumstances

https://www.theguardian.com/world/2021/nov/05/belgian-police...

https://www.theguardian.com/world/2020/aug/20/belgian-police...

And this is just a random example country (with excellent beer), you will find worse than this in most/all European countries, especially in the last 10 years or so.

This is a long post about someone who has very obviously just gotten into politics. It is good for people to try and see how to impart change. Here are some constructive critical questions for the author:

1. Why no mention of No Tech For Apartheid or Google Workers United, who have been doing similar work for years?

2. What about all of the other police, DHS, and military contracts Google has been a part of? Did this problem really just start with the second (not even the first!) Trump presidency?

3. What does a focus on exclusively those at the top levels of a hierarchy, with minimal focus on incentive structures and wider systems, say about your theory of change? Was there a power analysis done, or was it assumed that "big title" = "powerful"?

Side Note: Incredibly insulting of James Dean to say email 3 CEOs.

It is honorable but ultimately hasn’t democracy spoken wrt the issues mentioned in the post?

For better or worse millions of Americans voted for the guy doing the deportations.

I also find it difficult to reconcile not using AI for weapons. If war is inevitable AI presumably would at least ensure you are on target.

"You're absolutely right, that wasn't a military target—it was actually a girls school. It won't happen again!"

I realize that's not a great argument and was definitely tongue-in-cheek, but given there's still a lot of debate about the accuracy of AI for far more mundane tasks, my personal perspective is that until we have LLMs and such that are truly, demonstrably far more accurate than humans, with true reasoning and judgement capabilities, they don't belong where lives are at stake.

I wouldn't want an LLM-underpinned machine running anesthesia during a surgery; why would I want an LLM-underpinned military apparatus that is deciding the lives of far more? I wouldn't, not in their current state.

In a hypothetical future where we truly trust incredibly smart AIs or LLMs or whatever "smart" technology it is for driving weaponry, okay - if it's truly necessary; I abhor war and the death and destruction wrought by it.

In my mind, though - even if we get to that future where there's some vastly superior technology to the LLMs we have today, which can judge and reason, then I'll have a bunch of other questions, like understanding the motivations of said technology, because I suspect it'll be something much closer to AGI, and that opens a whole separate can of philosophical worms.

Mainly I don't want them because I don't want a machine instantly making life or death judgements. It might be fine if it mostly informs a human who ultimately pulls the trigger, but having that failsafe is important in keeping machines from going on killing sprees.

I also want people to be held accountable when they do unjustified killings. AI weapons make it FAR too easy to simply pass off a killing as a "woopsie doodle." It's just not acceptable to say "The algorithm made a mistake, version 23 will do better".

I don't have a problem with the AI providing additional information to it's user, but when that's incorporated into a weapon it's a short distance from that to completely automating the killing.

That's why I'm completely against AI weapons.

>Just out of curiosity though why do people not want smart AI weapons?

When the decision to kill another human being is made that should be in the hands of a directly accountable other human being, not an unaccountable machine developed in the basement of a private corporation.

And mines, both dumb and smart, in particular anti-personell mines are banned by the Ottawa treaty ratified by 162 countries. It's exactly the autonomous and fundamentally uncontrollable nature of mines, not just that they're dumb, that has produced countless of casualties long after wars were over. Can you tell me that millions of autonomous loitering munitions are not going to end up exactly like those mines still blowing legs off people decades after conflicts are over? And who is responsible then?

They aren’t training smart AI weapons to discriminate between civilians and soldiers. Even so, what happens when the soldiers are disguised as civilians? Or when the enemy forces civilians to serve or be a decoy?

Because it's going to be used in any possible way, and certainly not for the good of the people, as imply in your "naive" false dichotomy.

That doesn't address the supply chain risk designation? Or the ban on all work with claude for contractors, incl. non-government work?

Even if they disagreed, this was clearly retaliatory for not capitulating. The government could have just decided to go with a different provider.

Michael is some marketing guy who used to work for Uber and now chose, voluntarily, to work for Trump - and gets to cosplay at being "Undersecretary for War". Why would anyone believe anything that such a person says?

And this "might happen on the 30 minute clock of an ICBM" is just another variation on the Ticking Timebomb "if you only knew what I know" forcing-function that gets deployed all the time by these brave warriors.

I cannot think of a less trustworthy source of information than Emil Michael on the All-In podcast. Why on Earth would you take his statements at face value? I grant that nobody in this situation is totally trustworthy, but if it's his word against Dario's I'm believing Dario 10 times out of 10, and it's astonishing and dismaying that anyone might do otherwise.

Keep in mind that Emil Michael's job at Uber was to slander and destroy their political enemies. What makes you think he hasn't continued exactly as before?

Wait. Do you trust what he said?

It is not about courage or confidence, it is more a calling, where few are called and that often turns out to be just one person, or so it seems, since everyone finds their excuse to not take a stand against the oppressor and for the oppressed.

The author will have experienced every logical fallacy and TurnTrout has done well to document his efforts, rather than just the crimes of his adversaries.

A calling of conscience is hard to articulate initially, it is conscientious objection, to the war machine and the system of finance that necessitates the empire that needs the war machine.

The author starts off with a high degree of authority in that he actually worked for Google Deepmind, however, nobody will listen to him, so he has to seek higher authority to carry the truth forward. But he could have gone to the pope, anyone short of Jesus Christ (freshly teleported back) and it will be a no, everywhere.

Logic and reason does not help when insanity and money have taken over, which tends to happen during wartime.

The conscientious objector may not believe in god, however, they will consider themselves answerable, 'vain' enough to care about their legacy. There is that desire not to be found out, doing the 'devil's work', generally in workplaces where everyone is glad to get on with the 'devil's work'.

Why do so few people make a stand? Why didn't hordes of five-eyes people walk out the day after Snowden did his good deeds?

It varies by individual, however, the author is vegan, which means he has already 'dared to be Daniel, dared to stand alone', albeit only in a lunch queue in a meat eating world. This requires living according to principles, and serves as target practice for war-time conscientious objection.

Also important is a certain level of independence, the guy with a mortgage and a couple of kids, underwater on the car, with maxed out credit cards cannot conscientiously object. Nor can the guy counting down the last few years to retirement, and then the very young lack the articulation.

Only a few have the 'warchest' to embark on an open-ended single-person campaign to confront power with the truth. We owe a lot to these people, particularly the 'failed whistleblowers' that don't make it to 'whistleblower' status, because the media then makes the story about their situation, not what they conscientiously objected to. Props and respect to the author for documenting his journey and taking the first step.

Note the 'first step' is classic 'hero's journey', where the call is initially rejected, but then a journey into the extraordinary world is made, where the challenge is to bring back what is good from the extraordinary world to the ordinary world, for the benefit of all.

Working anywhere requires ethical compromise… a 2 person company can disagree on ethical questions

Competitors often wish they didn't have to compete.

For example, a cartel is "a group of independent market participants who collaborate with each other and avoid competition in order to improve profits and dominate the market"

It's something that could happen in this world, through the collective action of employees, that corporate strategy can be changed.

apple and google are regularly on the same side against eu regulations

On the upside, I guess maybe DeepMind is hiring.

... I'm sure I'll get flamed into oblivion for this but it's weird to me how the zeitgeist is anti-colonialism but also against enforcing borders and national sovereignty. I guess maybe they are okay with it unless you're a western nation. Whatever, there's no room for nuanced opinions anymore in modern online discourse.

I'll admit to being a weird outlier on this. I may not like what certain parts of the government are doing but I'd go work for the Voldemort companies in a heartbeat -- they just require in-office and aren't anywhere near me. I'd rather my nation develop the best technologies than let other nations do it.

Just look at how far behind the eurozone is by not making the right investments.

its more like: good luck finding clients who don't kill anyone somewhere deep inside their supply chain.

Agreed, I feel like the people who say no to AI weapons haven't actually presented a real argument (that I have heard) besides terminator bad

When folks who live around me vote for things I find unethical, then I don't feel bound by those decisions.

Also, your position on weapons development is premised on the idea that at lease some of the folks developing the weapons are on your side and always will be.

That might not be true at all.

AI helps people scale. If being "on target" means that you find people when they're eating dinner so you can kill their entire family, then using AI to scale that up is a terrible thing. That's exactly what Israel did with Lavender. They literally had a kill list called "Where's daddy?".

https://news.ycombinator.com/item?id=39918245

>AI presumably would at least ensure you are on target.

That is a big presumption.

> For better or worse millions of Americans voted for the guy doing the deportations.

Uh, the idea of democracy is not that voting reveals universal preferences or something. It is ok to disagree with whoever gets elected, and continue working towards an alternative you believe to be better. In fact, democracy depends on that; otherwise, why not have a single election with no term limits? There is supposed to be ongoing difference of opinion.

> I also find it difficult to reconcile not using AI for weapons.

You can't reduce "AI for autonomous lethal weapons" to "AI for weapons".

This idea that once a party gain a little more votes, people who oppose it should cease to exist and cease push for own values ... somehow applies only to the right.

When center or left wins, somehow, magically, same logic dont apply.

It does appear to be anti-democratic given everyone knew Trump’s platform when he was voted in. However the great unifying fact of politics is that everyone only really believes in democracy when they win.