OpenAI CEO Sam Altman Defends Pentagon Work to Staff | Hacker News Reader

OpenAI CEO Sam Altman Defends Pentagon Work to Staff

wsj.com

85 pointscdrnsf91 comments

It's not quite as bad as working for X, xAI, or Tesla, but engineers continuing to work for these companies are taking hits to their reputation as far as I'm concerned. Like, if I see it on a resume beyond a certain date, I'm not considering them types of reputational damage.

I've seen people saying that about Meta/Facebook for a decade, but I still don't see any tangible damage to former employee's ability to get jobs. The OpenAI situation seems much closer to FB-scale politics than X though.

I mostly agree with you re: Meta/Facebook except that things are becoming a lot more politically volatile than they have been in the past. Generally, I think that most people believe that the more intelligent you are, the more empathetic you are, so at some point if your evil company is doing big destructive evils, the smartest engineers will probably bail.

Depending on jurisdiction, it’s maybe not that smart to hop on the internet and write “if I see it on a resume beyond a certain date, I'm not considering them” and “things are becoming a lot more politically volatile” either.

"Generally, I think that most people believe that the more intelligent you are, the more empathetic you are..."

Okay, you might need to re-evaluate the life lessons you seem to have selectively taught yourself. This is base line culture war 'You must be mentally deficient if you don't align with what I deem to be empathetic right now or I don't think you're nice enough' type stuff.

High school type shit.

https://www.sciencedirect.com/science/article/abs/pii/S01602...

A study set out to investigate why the stupids lack empathy? Where did the premise for this bullshit come from? Haha

I think Amazon is a better example. It's a thing that some companies prefer not to hire engineers from Amazon because of the culture they bring. Whether you agree with it or not, Amazon has a reputation for a toxic culture and that sort of thing can ruin a smaller or medium size company if it seeps in.

OP wasn't talking about culture, he was talking about discriminating due to differences in political opinions, very different.

*ethical opinions

> but I still don't see any tangible damage to former employee's ability to get jobs

I have. I've also seen it happen for Uber, for someone who worked on the god mode project that went viral for being used at holiday parties.

I assume you work at some small startup where you get to dictate who you will hire based on your interpretation of what a candidate’s past work history tells you about their morals/ethics. But that shit won’t fly if you are interviewing at other large companies. You can’t reject someone just because they have OpenAI on their resume. In fact I have never heard of any FAANG company ever blacklisting candidates from some other company. So you rejecting someone is not going to move the needle much. They can leave OpenAI whenever they want and Zuck will offer them 8-9 figure pay packages :)

Oh, you absolutely can; maybe not as a matter of formal policy, but if you are a hiring manager or a member of an interview team, you have wide latitude to have concerns about nearly anything (legal) about a candidate. And also, even if you don't use that explicitly, it can affect your judgement of them when discussing them to an interview committee.

lol absolutely none of these things will fly on an interview feedback in any big company

> even if you don't use that explicitly, it can affect your judgement of them when discussing them to an interview committee

my friend that's literally unethical.

Bad news, Voltaire didn't actually say that thing about 'defending your speech to the death.'

i'm not defending anyone's speech - i'm saying just because other people are assholes doesn't mean i should become one.

Expressing qualms about hiring somebody who has voluntarily demonstrated a lack of ethics isn't 'being an asshole'. In fact, it's usually a good business decision and always an ethically-sound one.

Here's a sample question, feel free to use:

"If somebody offered you money to build the brains for fully-autonomous killbots PLUS mass domestic surveillance of the civilian population, would you accept the offer?"

"Yes" => They are unethical => No-hire.

"No" => They did do that, and thus are a liar, and thus are unethical => No-hire.

> absolutely none of these things will fly on an interview feedback in any big company

Sure, you never write "no hire because they worked at Palantir". You write "candidate didn't ask clarifying questions about {X} and jumped to answer {Y} which is not what I expect from a candidate of this level, no hire".

....this assumes that anyone at all reads your detailed notes if you submit an initial rating of "no hire", and I have very little evidence from my interviewing career across multiple companies to believe that's the case..

yea totally - in response to someone else's unethical choices i myself will choose to be unethical? because two wrongs make a right? correct?

> this assumes that anyone at all reads your detailed notes if you submit an initial rating of "no hire"

the director of my org (inside of FAANG) reads all of our interview feedback if we make an offer.

My experience at FAANG is that if you get a "no hire" vote, then unless several other reviewers are "strong hire" the candidate doesn't even make it to a committee and is already rejected - and this was before 2022.

that's not what i said - i said "if we hire, then he reads".

And my point is that if you write "no hire" than the candidate doesn't proceed and thus no one generally reads the feedback.

If you want to blatantly lie and hide your true reason for rejecting them by making up other stuff in the debrief notes, that would be possible. But at that point, you are the unethical person. You can technically do the same thing just because you wanted to discriminate based on race, sex, etc (that would be both illegal as well as violation of corporate HR policies).

well I'm just glad to find out nobody has been discriminating based on race or sex since that would be unethical

are y'all really repeating over and over "they did it so i can do it too"?

This thread is simultaneously horrifying and hilarious at the same time.

Hilarious because onesociety2022 seems so earnest. Someone who is shocked at the idea that job search isn’t a pure meritocracy.

Horrifying because kuang_eleven points out just how easy it is to pass a qualified candidate if you want to.

The truth is somewhere in the middle…

Honestly, would be the complete opposite for me if I was an interviewer (thankfully I am not and never will). You don't want fussy employees.

So you want yes-men in your team?

No, you just dont want people that will start interrupting work, causing a ruckus, starts signing open letters, or randomly quits based on whatever blue sky post they read last.

not working at xAI makes you a fussy employee?

fussy is a strong word, didn't mean it like that, but it is the fastest way to communicate it.

Yeah I don't work for fussy employers either. (Sarcasm) the shit goes both ways brother.

Exactly my point, as long as you pick a lane you will be good on both sides it's not an advantage or a disadvantage.

I will correct myself: Your employer might not want fussy employees (ex: palantir).

What I'm trying to say is fuck that if you're doing unethical stuff as an employer, you got problems. If you looking for am employee that sacrifices his principals for money that makes the world a worse place to be in and to ask for those things puts the employers in the seat of being fussy for the wrong things.

This is so lame.

Don't judge employees for what their CEOs do bc they do not have a choice in the matter.

That resume you toss might be someone that needs to pay a mortgage, has a sick wife, or autistic kid that needs the insurance. Or it could come from an employee who genuinely disagrees with mission and quit, but its not like you would know or even care.

What if your CEO went politically rogue and started openly supporting Trump? Would you quit? I doubt that.

What a childish attitude. Get your politics TF out of the office and remember the fact that we live in a democracy where sometimes you do not get your way.

> What if your CEO went politically rogue and started openly supporting Trump? Would you quit? I doubt that.

Yes. I would. Trump is a uniquely dangerous president with uniquely unrestrained power.

I'll continue to write my representatives, publicly protest, boycott businesses and employers, and use all other legal levers of power that I can. My kids future cannot afford to let these crimes continue unchecked.

> Get your politics TF out of the office...

A large portion of our lives is spent at the office, with people we may not otherwise interact with. Former coworkers sharing their perspectives helped sow the seeds of my change in politics, both on individual issues and worldview.

It's childish to think life can be perfectly compartmentalized, like pre-K learning stations.

In your understanding of the constitution, is the executive branch subordinate to the other two branches?

In my view the judicial and legislative branches should be no less powerful than the executive.

If there is truly an impasse the legislative branch should win because it is the most diverse and direct representation of the people.

I also have other crazy ideas like money isn't speech, ranked choice voting, same-day primaries, outlawing private financing of campaigns, bribery is accepting compensation even if before the action being bought, abolishing the electoral college and the US Senate, and no one is above investigation or the law--not even in their official capacity.

Insufferable.

Consider that more than 50% of your fellow American citizens voted for the man you hate, and you will have to work with them no matter what. That's called "democracy."

Its childish to avoid that truth.

People can elect racists. rapists, conmen, human traffickers, pedophiles, and even dictators. Doesn't mean I have to work with them or accept any of those as normal.

Resisting tyranny is a core tenant of democracy.

> People can elect racists. rapists, conmen, human traffickers, pedophiles, and even dictators. Doesn't mean I have to work with them or accept any of those as normal.

Trump is no saint, but you are turning him into a cartoonish villain. That never stopped democrats from taking his donation checks, did it?

> Resisting tyranny is a core tenant of democracy.

No, the core tenant of democracy is everyone gets a vote and you agree to live with the outcome, bc you get a vote too. And if you do not agree with the outcome, then you are the tyrant.

> Trump is no saint, but you are turning him into a cartoonish villain.

I didn't claim Trump is any of those things yet. Though courts would agree that he's a racist (in his tenant policies, never mind his retoric), a felon (financial crimes), a fraud (sending nonprofit money to Bondi's campaign), obstructs the biggest pedo investigation ever (redacting even "don't"), and the 14th amendment would disqualify him from federal office--if he hadn't packed SCOTUS (who rule for him 90+% vs 90+% against in lower courts).

> That never stopped democrats from taking his donation checks, did it?

Democrats have plenty of blood on their hands. At least some of them continue to try to reform campaign financing. Modern Republicans seem to relish in the money laundering aspects, especially Trump.

> And if you do not agree with the outcome, then you are the tyrant.

You cannot have democracy under tyranny. And if a country elects an aspiring dictator (who earlier lead a mob to overthrow an election result, sent fake electors, demanded votes be "found", demanded vote counting be stopped while he was ahead, etc) then I suppose it ceases to be a democracy thereafter.

Regardless, I'm not advocating illegal acts, despite the fact that Trump commits them on the regular. I'm advocating the exercise of freedom of speech, expression, and association in opposition to a (hopefully term limited) tyranny.

>> What if your CEO went politically rogue and started openly supporting Trump? Would you quit? I doubt that.

> Yes. I would. Trump is a uniquely dangerous president with uniquely unrestrained power.

Even if it meant you could lose your house? Even if you were 6 months pregnant (meaning it could put your health insurance in jeopardy at a pretty critical time, as well as your parental leave)? Even if you quitting would accomplish absolutely nothing?

Any maybe you're such a zealot that you'd rather end up homeless than working for a company managed by a CEO that supports Trump. But don't punish other people for not being willing to do the same. Other people have other priorities, and other situations going on in their life, and that's fine.

People said the same thing with the Nazi Regime. Im sorry at some point, you have to stop supporting a system that is destroying peoples lives. You cannot sit idly by and watch it happen because you need to pay bills.

Sorry, but if you continue to work for a company that develops mass surveillance techniques that will directly result in innocent people getting their way of life unjustly and unlawfully ripped away from them, then I think you suck and I don't know how not to think that.

If you work at Palantir, your reputation is forever tainted in my mind.

You should not deserve a job for the rest of your life. You should be homeless and honestly, maybe end your life.

If your CEO supports Trump, that probably isn't the best thing either, although this is 1 million times better than being the tool used to break every laws, developing killing and spying machines.

Alex Karp has publicly supported Hillary Clinton and Kamala Harris. His company did work with the government before Trump, and you probably never had a problem with it until now.

Palantir always worked with the military, but I doubt you cared when they helped drone people during the Obama administration.

I actually don't really have the biggest problem with Trump. I voted for him.

I do have a problem with surveillance.

> You should not deserve a job for the rest of your life. You should be homeless and honestly, maybe end your life....I actually don't really have the biggest problem with Trump. I voted for him.

Some people here would say you should not deserve a job, be homeless, and maybe end your life bc you voted for Trump. How do you feel about that?

I don't get it. Interesting problems are interesting problems. Granted, I don't think you see those at X, but xAI, Tesla or Meta? Sure. I am not even arguing 'money' or 'man gotta eat', but just seems like such an arbitrary thing to flag.. especially since it won't be documented anywhere ( 'we are anti-Tesla house' banner on main page for example).

It is silly.

It ain't silly. You're working for Hitler

Interesting problems don't exist in a vacuum. I'm sure it was an interesting problem to figure out how to track people who opted out of tracking, how to build gas chambers, how to add lead to gasoline, doesn't mean one should choose to solve them.

> how to add lead to gasoline

Your other examples line up, but this wasn't a problem being solved - engine knock was a problem being solved.

If we are going to claim to be software "engineers" we have ethical obligations. That means that you don't just do whatever the person signing your paycheck says, you raise objections and refuse to do things that will cause harm.

In real engineering disciplines, engineers sign their name to key decisions and if people get hurt someone loses their license and their right to do engineering work. The world would be a better place if software worked that way, although it'd be harder for a bunch of sociopaths to become billionaires.

I guess you could consider "literally building Skynet" to be an interesting problem, but you have got to have some pieces missing if "interesting" is the most appropriate assessment of that work.

The backlash is "painful"? Maybe don't make a moralizing tweet about your principles only to change them three hours later. It comes off as opportunistic grifting.

I don't blame the OpenAI staff (and as far as I am aware most people don't). Most of us end up working for assholes if you go far enough up the chain, but it's different when the CEO tries to earn social credit by having his principles, only to seize on an opportunity and just ignore those principles. He can say "oh well they pinky swore they wouldn't abuse this or redefine laws to say what they're doing is 'lawful'", but I personally would have trouble trusting the words of a convicted fraudster lolcow that we decided to elect as president and an alcoholic Fox News host. I guess that makes me a "radical leftist" though, I'm sure that the 10000 IQ people always trust convicted criminals.

I'm sure Sam Altman will make his money, and I'm sure that OpenAI will continue to take over the world like before, but I don't have to fund it myself, hence why I canceled my ChatGPT Plus and signed up for Claude. I'm sure that the CEO for that company will be a douche eventually too, but at least as of right now he seems to have a shrapnel of integrity.

I can't read most of the article because most of the common archiving sites don't appear to work.

"I'm sure that OpenAI will continue to take over the world like before"

Each day that goes by Im more convinced OAI will not be a healthy going-concern without government help, which most likely will not be granted.

I dunno. They were valued last week at $700+ billion weren't they? When you have that kind of capital available I'm not entirely sure how possible it even is to go bankrupt.

Regardless, my point is that one dude canceling his $20/month subscription probably isn't going to affect anything, but it's basically all I can do.

"They were valued last week at $700+ billion weren't they?" So?

The money they have available is what is on the balance sheet, which they are burning right-through whilst facing immense competition and never-ending reinvestment, given that Google will carry on doing so. Cash flows from operations is a big fat negative.

I see Google first killing OAI, then eventually doing the same to Anthropic, once they figure out a suite of products that truly revolutionises the work of a sofware engineer beyond just talking to a chat interface, and bundle it into their existing offerings for enterprise.

I guess I just feel like when you're worth that much, you can be unprofitable for a very long time before it catches up with you. That's my perspective anyway, I could be wrong.

I agree that if anyone is going to kill OpenAI it's likely Google. They have even more funding and already have giant training indexes for search that they could likely leverage to improve their models in a way that OpenAI can't.

Thing is, that valuation is kind of based on vibes. If people get less enthusiastic about them (or just the sector in general), that valuation will go down. Their ability to raise money off it is limited.

Yeah it takes awhile for investors to start actually valuing things, they usually 'price' them which is mainly driven by mood and momentum.

Given that pricing is the name of the game, all should be wary of what interests founders and their employees who are paid via stock based compensation...

The money they raised is basically already spent.

On DRAM that hasn't even been fabbed yet. :(

Trying to create shortages etc is just typical Scam Altman behaviour and really speaks to how vulnerable he feels against Google's might.

There was a point where Google's existence was questioned, but they've been working away quietly and they'll win the long game. Altman viewed OAI as a way to reduce Google's AI dominance - I think when we look back in history it'll turn out he made it worse. Google wasn't all that interested in releasing LLMs out into the wild.

I tend to think you're right about Google. Just for the simple reason that webmasters everywhere try their best to keep OpenAI's crawlers out, while simultaneously trying their best to attract Google's.

The "Backlash" he is referring to is people canceling their OpenAI accounts and going to Claude. And by "Painful" he means less money. - Professional CEO Sam Speak Interpreter

I mean, I guess. It's not like they're profitable yet are they? At least not at the consumer level as far as I am aware. Me canceling my subscription might have saved them money.

Honestly I think these tech billionaires are very thin-skinned and they don't like people saying mean things about them, and I think a lot of them are completely unable to reconcile this simple fact: when you're a billionaire, you don't fucking get to have a normal life.

You didn't have to get billions of dollars of wealth. If you get into a situation like that, then yes your actions are going to be scrutinized more than a nobody like me. People are going to call you an asshole when you do asshole things more than when some random nobody does something assholey. You chose to be popular and powerful; if you don't like that then there's no law saying you can't get a regular 9-5 job like the rest of us.

Trickles often turn into streams and streams into rivers.

> don't make a moralizing tweet about your principles only > to change them three hours later.

I have no inside scoop, but it sure looks like this was all pre-arranged. Altman made a better offer to Hegseth/ Trump (or offered some other "inducement"), so Hegseth found this way to weasel out of the contract with Anthropic. I don't see how this all would have transpired that quickly otherwise. And of course the fact that three days later OpenAI reportedly got the same contingencies on its contract that were the supposed reason for cancelling Anthropic's contract just looks wrong.

Entirely possible, though I think it might have been that Anthropic wanted extra guarantees outside of what was "lawful".

The stuff that Altman mentioned seemed to indicate that they'll support the US government as long as the US government is following the law, ignoring the "if the president does it it's not illegal" mentality that this administration appears to be taking.

I'm choosing to wear "radical leftist" as a badge of honour, just by the words and actions of those who use that terminology for condemnation.

Stay the course (figurative) brother.

(This account may be permanently disavowed in the unlikely event I need to travel to the US)

I built a website that shows a timeline of recent events involving Anthropic, OpenAI, and the U.S. government.

Posted here: https://news.ycombinator.com/item?id=47195085

> Access is temporarily restricted

> We detected unusual activity from your device or network.

Anyone have an alternative link? The archive.* sites are also an endless captcha loop for me unfortunately. And no I am not using any VPN or CF DNS/etc.

I'm getting lots of that "We detected unusual activity from your device or network" lately. I even went through my ISP to change my IP address, to no effect. If there really is "unusual activity", I wish they'd be a bit more specific.

HN bans IPs for unusual activity too, luckily they have a human friendly way to get unbanned.

I usually skip WSJ articles because of that, but this worked when I checked:

https://smry.ai/proxy?url=https%3A%2F%2Fwww.wsj.com%2Ftech%2...

PS. I can't be sure if it's the whole article, not just some AI summary thou.

blaming open ai emps is like blaming current germans for ww2. that sort of collective moral guilt, sometimes even inherited, is simply unfair and stupid. i get that people want to undermine companies' support structure and that their dream is probably to guilt shame open ai's employees into quitting. but it feels like people trying to score points for their own agenda, claim the moral high ground, and label entire groups of people with some sort of guilt by association. its fake. its a game. and they know it.

I don't like the analogy anyway, but why 'current Germans'? Surely 'pre-enrolled rank and file of the contemporary military' would be more apt?

"IBM employee who built the software which effectuated the holocaust while knowing that was what it was being used for" is probably more apt.

But we’re not taking about a 80-year gap here (“blaming modern Germans”); we’re talking about people who are in the global top 5% of income and prestige choosing _today_ to contribute to these organizations.

If you believe that your labor is worth something — which I’m pretty sure this crowd does — by working for a given firm, you’re voting with the value of your time in support of what your employer does.

Which, to be clear, is 100% your choice! I’m not going to accuse anyone of being a “bad person” because they decide that stable, high-paying employment is more important than taking a particular ethical (or political) stand at work.

But it _is_ a choice that you make every day by showing up for work.

In my view this is even more relevant for tech workers who receive equity. If you’re a shareholder in addition to being an employee, you’re now voting _twice_ in favor of what management is doing, and benefitting directly from both pay and ownership.

fair point, and i do agree with you. i guess ive seen too many politicians recently with agenda's doing the moral high ground thing and im beginning to see that sort of conduct everywhere. thanks for your reasoned comment. it seems to be a growing problem or maybe im just more aware of it

I'm surprised this thread is only on page 5, despite 65 pts and 66 comments in 3 hours.

What other factors go into listing position?

Reply on news.ycombinator.com