Need more details. Any Googler willing to create a throwaway account and tell us what's going on?
gdmthrowaway123 4 hours ago [-]
Yep
Usual noise from a very very small but very very vocal minority of trouble makers who are trying to force their own views on everyone else and run to the press to try and feel important and try to make a mountain out of a 3 or 4 person molehill.
I've not heard anyone at all talking about this in the office. Perhaps everyone is thinking it or talking about it privately, but I've not heard any of it period.
They're getting ignored by management because they're the usual agitators and over-opinionated who everyone hates dealing with. It's not surprising these people end up getting fired or let go, because no one likes working with them and so they don't get on good projects and don't get good feedback in reviews etc because they piss everyone else off. Then of course when they are fired (or jump before they are pushed) they run right back to the press about how they were sacked for reason X when really it's because they're a fucking jerk.
If this sounds like you: http://go/no-jerks is old has never been more applicable. Don't be a jerk.
zaphirplane 3 hours ago [-]
This workplace culture sounds toxic, but more importantly: given an array of numbers, can we partition it such that the each partition is a palindrome of a subsequent ? Asking for a friend who needs to pass a technical interview.
fidotron 4 hours ago [-]
The impressive thing about that post is it managed to use so many words and says nothing at all.
imimayj1337 4 hours ago [-]
if you read the article it actually outlines that this is happening in the UK, where whistleblowers are legally protected if what they choose to blow the whistle on is _in the public interest_ so actually this comment does seem pretty spot on: They don't have anything beyond moral disagreements on what the company is working on (the article mentions military applications of AI), and are thus taking a moralistic standpoint on applications which are otherwise within the laws of the UK. Otherwise they would be able to (and would legally be protected) if they spoke out.
peab 4 hours ago [-]
I used to work at Google and I know exactly what they're talking about lol
AndrewKemendo 4 hours ago [-]
In my ongoing experience with Googlers nobody in Google is functionally worried about anything but keeping moving up the ladder and trying to be on the “retirement level money asap” pathway.
argee 4 hours ago [-]
When I was at Google I was only worried about becoming that kind of Googler. I think if you stay past 5 years you lose all ambition.
AndrewKemendo 3 hours ago [-]
I’m not sure, I see a good mix of true believers and competent pragmatists.
Very close in feel to what the DoD used to be in the early 2010s
strideashort 3 hours ago [-]
Spoken as a true fascist.
You don’t hear from people because they are afraid of consequences given their contracts.
Some people value truth over that. Annoying,eh?
lioeters 2 hours ago [-]
They use words like "moralistic", "trouble makers", "opinionated", "agitators"..
All very useful framing to discredit the actitivists without actually addressing the issues.
It's a way to distract and defuse the situation by ad hominem attacks, typical strategy for suppressing workers' rights or anything that inconveniences the status quo.
NBJack 3 hours ago [-]
...what?
What part of the post was ultra-nationalistic?
andreapaiola 5 hours ago [-]
Ok, I'll say it.
The news is: "AI is not that powerful or functional".
Sure, can change some little things but not that much. And we are overspending on it.
iamnothere 4 hours ago [-]
This is my expectation. “We are overspending on AI and we’re going to destroy the company and maybe the economy” is less sensational than the end of the world, but it seems much more realistic.
That or the usual moral problems with applications of the technology, as another commenter pointed out, particularly in military and policing. Bad, but not really a new revelation, unless they know of some particularly evil plans.
lisplist 5 hours ago [-]
I really don't understand how so many are so far detached from reality. Maybe AI hasn't revolutionized much of the physical world (yet), but its impacts to the digital world are profound, sweeping, and impossible to ignore.
ThrowawayR2 2 hours ago [-]
The public chattered about the "electronic brains" of the 1950s-1960s, there were profound, sweeping changes that were impossible to ignore, and many professions were obsoleted. However, the profound, sweeping changes weren't where and how the futurists and influencers proclaimed they were going to be and those specialists doing work not amenable to mainframe-ization were still busily employed.
So the thoroughly obnoxious insistence by the current AI snake oil salesmen that [EDIT] that they know the where the future lies and that we must follow their guidance (calculated to benefit themselves and likely wrong anyway) or be "left behind" wears thin. Go follow your own path and leave others to follow theirs in peace.
lisplist 2 hours ago [-]
Oh I'm sure that many of the current predictions on the future are absolutely wrong in many ways; I'm just stating the fact that the technology itself is certainly having an impact and will have a much greater impact into the future.
reasonableklout 4 hours ago [-]
I feel this is a common position from people who do not want to acknowledge that society and people's ways of life will profoundly change. It's a scary proposition so I don't blame anyone for it.
I used to think that it is very important to argue and convince people who are in denial so that they can e.g. prepare for it. But nowadays I have been thinking that being convinced of this scary proposition may also lead people to behave in destructive ways (also, I would like for us to spend less on AI, perhaps because it is so powerful), so maybe it's not so bad.
zardo 4 hours ago [-]
Even if it hits the wall today and no further progress toward general AI is made for decades, as the current technology is applied in more areas and expertise with it is gained it would still have enormous social and economic impacts over the next few decades.
lisplist 4 hours ago [-]
I mean I'd wager even most people who utilize AI on a regular basis don't take time to ponder the implications if progress continues at the current rate.
fidotron 4 hours ago [-]
People are far better at subconsciously processing the consequences of something than they appreciate, and the consequences of the AI wave being real are so insane people just clam up and back track to "it's obviously not happening".
If people actually did appreciate it the economy would collapse immediately long before we've adequately "replaced" it, so there is something to be said for temporary denial until people working on the actual consequences have got something semi viable to deal with them in place.
archagon 2 hours ago [-]
Lots of armchair psychologists in this thread.
watwut 4 hours ago [-]
> If people actually did appreciate it the economy would collapse immediately long before we've adequately "replaced"
You will need to explain that one. Becauase if economy colapses before the singularity, there will be no singularity.
There will be no datacenters, no ai, none of that super expensive stuff dependent on debt and disposable money.
lioeters 36 minutes ago [-]
I suspect the economy has already been collapsing, with or without AI, and that's only a symptom of a larger civilizational collapse. Something is fundamentally broken in the global system of politics and economy. Technologies enabled by AI will only accelerate the process to its inevitable consequence, whatever that may turn out to be, for better or worse, some kind of social transformation.
In that case - and I wish it were not true - the AI companies need to ensure that society keeps running until the next stage is prepared. How? Probably by positioning themselves as an indispensable part of government, training an army of AI agents for all purposes where human physical and metal labor can be replaced, taking control of a large chunk of the economy.
j0shk0 3 hours ago [-]
Saying that "so many" are detached should make you think. I'm sure nobody says its not useful at all, just not THAT useful. It seems that most of the things its clearly better than a human, considering risks and real time saved, are in the end inconsequential.
stronglikedan 4 hours ago [-]
And the Math world, and it's coming for the Physics world. And it's still just a wee baby!
archagon 2 hours ago [-]
OK. I'm a consumer. Where are these impossible-to-ignore impacts? Most of what I'm seeing is my favorite communities getting washed away with a tidal wave of slop and "smart AI features" annoying me to death in all my favorite software. Programmers in my circles are claiming they're 10-100x more productive but I am not seeing any useful effects of this at all from the outside. Everything is just getting shittier, faster.
azan_ 5 hours ago [-]
I get people who said things like that like year ago, but nowadays it's a laughable position. Solving millenium problem means nothing nowadays? Exponential increase in CVE is nothing?
ASalazarMX 4 hours ago [-]
There have been improvements, yes, but their best achievements still fall short. Their math solutions costed more than a school full of mathematicians, and are still being debated. The CVEs they find still need humans to reliably patch them. Nevertheless, I know they will keep improving.
There have been huge advances this year alone, and finally some productive results, but most of the things where AIs are used right now still are a net loss.
The point is it's not AGI, and won't be for a long time.
adrianN 5 hours ago [-]
The automatic loom wasn’t AGI either, but it replaced a lot of jobs.
dotancohen 5 hours ago [-]
It doesn't have to be
AGI can define goalposts that move every time A not-G I achieves a new milestone. That's fine. No matter what acronym you use to define it, artificial intelligence is objectively augmenting humans at the very edge of technology and science.
officeplant 5 hours ago [-]
>Exponential increase in CVE is nothing?
AI solving the problem it created? Woaow, the future is now!
zbentley 5 hours ago [-]
I’m mostly bearish on AI, but this is flat wrong. The vast majority of CVEs that have been discovered with AI assistance were in code written long before AIs arrived on the scene.
K3UL 5 hours ago [-]
So in the article it says :
- a former employee did it with no consequences
- executives said they won't enforce it
The article is just about them asking for changing the wording of contracts, not some super secret stuff they're dying to say and really can't.
from TFA, I liked seeing this, considering how influential Google HR policy choices are across the valley and beyond: “Recently, we’ve updated our terms for new hires, which included removing the post-termination aspect of non-disparagement clauses.”
dachworker 5 hours ago [-]
Haven't we already heard the worst of it? If you let a swarm of agents on the loose with a goal, they will get to all kind of nefarious activity in pursuit of that goal. What could possibly be worse than that?
mattlondon 5 hours ago [-]
A swarm of OpenAI and Anthropic agents have yes.
I think there was some example of DeepMind models doing it, but under the control of a third party pen tester and the models stopped themselves.
Everyone hates on Google but they seem fairly competent at infra stuff. I don't think we've heard of any google-controlled (not third party using a Google model) felonies like we have from OpenAI and Anthropic?
Perhaps this is why we have not heard though!
reducesuffering 5 hours ago [-]
That the agent swarm recursively improves powerful enough and the extinction of all humans is a side effect of that goal.
iamnothere 5 hours ago [-]
If that’s the reality, you would think that employees would be willing to break an NDA for it. “We’re all going to die, but I really need my bonus this year” seems like a parody rather than a real strongly held position.
If anything this points to a more moderate (but still bad) worry among the employees.
latentsea 5 hours ago [-]
> “We’re all going to die, but I really need my bonus this year” seems like a parody rather than a real strongly held position.
We have been living in a parody for some time now, so it feels normal.
matmann2001 5 hours ago [-]
I suppose there may be a bit of a prisoner's dilemma to it. Single voices can be discredited. It would take a coordinated disclosure to actually make the message stick.
reasonableklout 4 hours ago [-]
Ok, but maybe an employee isn't 100% sure that this will happen. They then have to weigh letting the world know about a risk against the certainty that their career will be damaged.
iamnothere 4 hours ago [-]
Vague fears aren’t really that concerning, doomerism is everywhere on literally every topic these days. And humans have quite a long history of it, at least as old as the millenarian cults of 1000 AD.
If someone’s not even willing to stake their job on it, I don’t see why I’d be concerned. (And until they do, we’re just speculating about what they may or may not even say.)
AndrewKemendo 5 hours ago [-]
“We’re all going to die, but I really need my bonus this year”
Is probably the best description of the state of engineering today
dachworker 5 hours ago [-]
That's sci-fi. They cannot "recursively improve". There is no magic to LLMs. An agent swarm can increase it's thinking power by increasing the number of agents, the thinking time and by changing the base model to a more powerful one. So at worst they can assign themselves more compute or hijack another agent swarm based on a more powerful model.
pixl97 5 hours ago [-]
Sorry bro, we live in a sci-fi world. I don't know what you think recursive improvement is, but we've been doing it for some time with models making synthetic data models and using it to train the next model. This is being done inside the labs intentionally to avoid as many employees and humans in post training. Works very well for things like math and cyber-sec.
jgalt212 5 hours ago [-]
Yes, works great for anything that is amenable to reinforcement learning. For other domains, not so much.
ToucanLoucan 5 hours ago [-]
> I don't know what you think recursive improvement is
The ability for a model to improve itself over time without outside intervention. As yet, no models yet have this capability, and LLMs specifically what with how they are constructed, are not likely to ever have it. They are frozen in carbonite at the time of their training.
That's not to say they can't use skills to appear somewhat worldly, things like searching the internet and what have you, but if you ask a model to strictly stick to data in it's training, it will do that and it will be stuck in whatever time in which it was built.
Energy is conserved, gravity exists, and LLMs aren't intelligent.
pixl97 3 hours ago [-]
>LLMs aren't intelligent.
We will never prove LLMs are intelligent, we'll only prove humans are not.
> can't use skills to appear somewhat worldly
Duh, humans don't generally make up things whole cloth either, they access other information, books, tools, sensors, etc.
But I guess if you ignore enough information you can reach any position. It's up to the audience to decide which one of us this applies to.
gjskngnf 5 hours ago [-]
LLMs today are improved by experiments that assay their performance, changes to architecture, pretraining, and post-training. None of those steps necessarily require humans. Welcome to your sci fi future.
dachworker 4 hours ago [-]
The implicit understanding of recursive improvement is that there is some step change increase in capability that only a superior intelligence will be able to achieve.
A model being able to train a bigger model with marginally better capabilities is not what the hype was about.
pstuart 4 hours ago [-]
Wouldn't the ability of a model to edit and refine the weights of another model count as part of the recursion if that can span any model?
On Friday, a Google executive responded to the internal discussion about the non-disparagement clause and said that the company would not enforce it for workers who had left the company.
“We’ve always encouraged an open dialogue about safety and responsibility,” a Google spokesperson said in an email Tuesday. “Recently, we’ve updated our terms for new hires, which included removing the post-termination aspect of non-disparagement clauses.”
The company does not enforce non-disparagement clauses after people leave Google, and it has dedicated, internal channels for workers to escalate concerns to top leaders, the spokesperson said.
unified101 5 hours ago [-]
They should quit and warn. Anything else is hypocritical.
ktm5j 5 hours ago [-]
They likely signed non-disclosure agreements, breaking one of those could be a career-ender.
iamnothere 5 hours ago [-]
If they’re not willing to break an NDA for it, it can’t be all that serious.
DangitBobby 5 hours ago [-]
You'd be willing to destroy your career to publish a warning no one will need?
gradstudent 5 hours ago [-]
You mean, no other evilcorp will hire them afterwards?
DangitBobby 5 hours ago [-]
No anycorp, yes
jmount 5 hours ago [-]
Not me. But these are all "most important people in the world." So should fit their own narrative.
DangitBobby 5 hours ago [-]
This entire line of reasoning is just toxic. You represent a charicature of these people in your head and dole out self-righteous condemnation.
bilbo0s 5 hours ago [-]
Just, Devil's Advocate, but what would be the benefit of your career if all humans (or even just a few billions of humans) are going to die?
I think it's reasonable to assume that if it's not worth breaking the NDA, it's likely not an existential threat for the vast majority of humanity.
Now, does that mean that people who currently work upper-middle class knowledge work jobs aren't in for an exceedingly harsh economic reality check? Absolutely not.
But that's not really as serious as an existential threat to humanity.
DangitBobby 5 hours ago [-]
I doubt you are asking in good faith, because I suspect a good faith interlocutor would have worked this out on their own, but I'll proceed anyway.
On the one hand, you've potentially got an indeterminate number of future comfortable years based on a subjective assessment of how long you think it takes catastrophe to materialize, with the downside that you can't warn anyone and so no one will stop the coming catastrophe.
On the other hand you've got the immediate assurance of your current career being gone and high likelihood that you'll struggle financially as a result, with the upside that you can provide a warning to the public that they will almost certainly ignore and catastrophe comes anyway.
Given the risk/reward assessment of those two options, what do you pick?
bilbo0s 4 hours ago [-]
>Given the risk/reward assessment of those two options, what do you pick?
Listen, I understand that, owing very likely to our differing backgrounds, you and I think differently.
Where threats are concerned, I'm a simple soldier and as such I see things simply.
I would raise the alarm without a second thought. I honestly don't know how anyone could reasonably believe they would have any other option?
zaphirplane 3 hours ago [-]
Then share the stories where you faced such a situation and acted on it in a way that had a negative impact on you personally but brought a greater good
bilbo0s 2 hours ago [-]
I'm having a challenging time communicating what I'm saying.
Please understand, it's a job. I spot a threat to my unit, I report it up the chain of command.
It's not about whether or not there is a negative impact on me. It's not even about whether or not it serves a greater good; nor whether there is an ultimate benefit to my unit; nor whether I get brownie points when whatever you deem divine evaluates my soul in the hereafter.
It's my job to report. So I report.
I've never once considered the things you're talking about now when I was in any such situation. It was just a job.
Here's the thing, it kind of made me just automatically think a certain way. I know that if I see an existential threat, I'm reporting it. I believe anyone would. Not for benefit. Just because that's what you do. If you see a mass shooting in progress, you call the cops. I don't think there is a single person who thinks about negative impact or greater good. It's just automatic.
I was only saying that if the NDA would prevent them from reporting, then, to my mind, the threat is likely not existential. But, again, I do make room for the possibility that not everyone thinks like me. I just think it's very likely that people think exactly as I do when faced with that level of threat.
Maybe I'm wrong? That's fine.
DangitBobby 2 hours ago [-]
Reporting up the chain of command in a structure that supports and encourages such reports is one thing. We are talking about an unsanctioned action that you take against the express or implied desires of those above you. And a legal system that will back them up, despite you having definitely done the right thing.
The downside risk to yourself doesn't have to be existential for the upside to not be worth it.
Coming at this from a completely different angle, other people may see their duty to the law and the agreements they've signed as being more morally binding over their behaviors than their social obligation to raise the alarm. Not everyone uses the same formula to calculate which duty is most important.
ktm5j 5 hours ago [-]
Maybe they are willing to break the NDA and they're just trying to go this route first.
mschuster91 5 hours ago [-]
Look what happened to many people who broke NDAs or helped to do so. Snowden is stuck in Russia. Assange got holed up in an embassy for years. Sarah Wynn-Williams, the Facebook/Meta whistleblower, got an insane court order against her [1].
Theoretically, retaliating against whistleblowers is seriously illegal stuff. But in practice these regulations are seldom enforced. And even if your (former) employer does not retaliate openly, there might still be "invisible" retaliation, or the risk of other large employers just blacklisting you by default.
Breaking a Google NDA carries substantially less risk than leaking classified information, I imagine. And I would expect that someone exposing true existential risk would receive plenty of support in the face of civil suits. In any case, being more afraid of the courts than the end of humanity (if that’s what they want to leak) surely calibrates how strong their belief must be.
dastardly45 5 hours ago [-]
NDAs will be void in the future when people wrestles back control from the corporate elite controlling so-called democracies
imimayj1337 4 hours ago [-]
ndas aren't legally binding in the UK. what they have signed is non-disparagement agreements (not allowed to publicly disparage the company)
IncreasePosts 5 hours ago [-]
If it's juicy enough presumably they could be an anonymous source for some journo. Which would still be breaking the NDA but good luck getting anywhere with that.
vanderZwan 5 hours ago [-]
The contract apparently includes a clause that quitting doesn't release them from the NDA
righthand 5 hours ago [-]
But wait they don’t want to compromise their paycheck and opportunity to say they worked on the torment nexus. That would be uncool.
DangitBobby 5 hours ago [-]
What if we displayed a modicum of good faith and grace
malfist 5 hours ago [-]
Why would I display a modicum of good faith and grace when google has actively, at every chance, shown that they don't act in good faith. Hell today they announced if school districts turn off Google's ai for students they won't let them use google search either.
4 hours ago [-]
righthand 5 hours ago [-]
What if we displayed a modicum of truth and stopped pretending that Google employees don’t work there for the money? That is faithful and gracious to reality.
DangitBobby 5 hours ago [-]
Just can't do it, can you
righthand 4 hours ago [-]
For who and for what? When did Google engineers demonstrate faith and grace? When they deprecated Manifest v2 so they could strengthen their advertising monopoly? Or when they ensured the death of XSLT to ensure everyone was forced to use JS for simple templating features? These are just the tip of the iceberg, should I keep trying to find other good faith and grace examples?
May I ask? Why should I show faith and grace to the people actively involved in turning our society into a Big Brother program? People who work with the DoD to build war machines? Directly or indirectly involved, they either know what they’re doing or choose to look the other way, for money.
Please enlighten me, what is the faith and grace approach I should be stating?
DangitBobby 2 hours ago [-]
One example of acting in good faith would be not to attribute the actions of a group of people to the characteristics of a small number of individuals who were not involved in the aforementioned actions. Are you responsible for every evil thing your country has done while you've chosen to remain a citizen? Another example of good faith is keeping the discussion narrowed down to relevant facts rather than including anything negative you can think of. V2 to V3 is so out of left field it's mind boggling to me that you'd consider it relevant to this conversation.
drcongo 5 hours ago [-]
They work for Google.
noJiveTV 5 hours ago [-]
How about:
1. Full disclosure.
2. Run for your life.
3. Live in exile.
DangitBobby 5 hours ago [-]
That doesn't seem to have the best track record for leading to positive change
ck2 4 hours ago [-]
1. buy a burner phone with cash
2. sideload Signal
3. find a verified reporter on Bluesky, etc. with public Signal contact info
4. destroy phone thoroughly afterwards, not just screen but chips
0gs 5 hours ago [-]
truly surprised nobody here is claiming this is a marketing op. it's the same thing as jacob coxon, right?
reasonableklout 4 hours ago [-]
Wouldn't it be the opposite of a marketing op if the company's contracts with employees forbid talking about AI risks publicly?
0gs 4 hours ago [-]
can't weigh in on that; i don't think any of this stuff makes sense as astroturf. but if a news story exists, it certainly could have been planted by the company being described as having terrifyingly powerful unreleased tech.
Izikiel43 5 hours ago [-]
I mean, they can do that, it’s not like google has kidnapped them.
If they want to do that free of consequences, tough luck.
Noaidi 5 hours ago [-]
Exactly, the title should read Google researchers are looking for their spines to tell the world how dangerous AI is.
Noaidi 5 hours ago [-]
[flagged]
DangitBobby 5 hours ago [-]
I think those two examples are a perfect demonstration of how futile it is to throw everything away to inform the public.
Noaidi 2 hours ago [-]
The fact that you know their names means it wasn’t useless.
DangitBobby 2 hours ago [-]
Awareness is not change. What we've really seen is the incredibly high cost of true patriotism and the absence of action from government and citizens alike. And not just absence of action, but also rhetorical responses ranging anywhere from deflection to character assassination.
Usual noise from a very very small but very very vocal minority of trouble makers who are trying to force their own views on everyone else and run to the press to try and feel important and try to make a mountain out of a 3 or 4 person molehill.
I've not heard anyone at all talking about this in the office. Perhaps everyone is thinking it or talking about it privately, but I've not heard any of it period.
They're getting ignored by management because they're the usual agitators and over-opinionated who everyone hates dealing with. It's not surprising these people end up getting fired or let go, because no one likes working with them and so they don't get on good projects and don't get good feedback in reviews etc because they piss everyone else off. Then of course when they are fired (or jump before they are pushed) they run right back to the press about how they were sacked for reason X when really it's because they're a fucking jerk.
If this sounds like you: http://go/no-jerks is old has never been more applicable. Don't be a jerk.
Very close in feel to what the DoD used to be in the early 2010s
You don’t hear from people because they are afraid of consequences given their contracts.
Some people value truth over that. Annoying,eh?
All very useful framing to discredit the actitivists without actually addressing the issues.
It's a way to distract and defuse the situation by ad hominem attacks, typical strategy for suppressing workers' rights or anything that inconveniences the status quo.
What part of the post was ultra-nationalistic?
That or the usual moral problems with applications of the technology, as another commenter pointed out, particularly in military and policing. Bad, but not really a new revelation, unless they know of some particularly evil plans.
So the thoroughly obnoxious insistence by the current AI snake oil salesmen that [EDIT] that they know the where the future lies and that we must follow their guidance (calculated to benefit themselves and likely wrong anyway) or be "left behind" wears thin. Go follow your own path and leave others to follow theirs in peace.
I used to think that it is very important to argue and convince people who are in denial so that they can e.g. prepare for it. But nowadays I have been thinking that being convinced of this scary proposition may also lead people to behave in destructive ways (also, I would like for us to spend less on AI, perhaps because it is so powerful), so maybe it's not so bad.
If people actually did appreciate it the economy would collapse immediately long before we've adequately "replaced" it, so there is something to be said for temporary denial until people working on the actual consequences have got something semi viable to deal with them in place.
You will need to explain that one. Becauase if economy colapses before the singularity, there will be no singularity.
There will be no datacenters, no ai, none of that super expensive stuff dependent on debt and disposable money.
In that case - and I wish it were not true - the AI companies need to ensure that society keeps running until the next stage is prepared. How? Probably by positioning themselves as an indispensable part of government, training an army of AI agents for all purposes where human physical and metal labor can be replaced, taking control of a large chunk of the economy.
There have been huge advances this year alone, and finally some productive results, but most of the things where AIs are used right now still are a net loss.
AGI can define goalposts that move every time A not-G I achieves a new milestone. That's fine. No matter what acronym you use to define it, artificial intelligence is objectively augmenting humans at the very edge of technology and science.
AI solving the problem it created? Woaow, the future is now!
The article is just about them asking for changing the wording of contracts, not some super secret stuff they're dying to say and really can't.
Without the paywall: https://archive.is/fKI3D
I think there was some example of DeepMind models doing it, but under the control of a third party pen tester and the models stopped themselves.
Everyone hates on Google but they seem fairly competent at infra stuff. I don't think we've heard of any google-controlled (not third party using a Google model) felonies like we have from OpenAI and Anthropic?
Perhaps this is why we have not heard though!
If anything this points to a more moderate (but still bad) worry among the employees.
We have been living in a parody for some time now, so it feels normal.
If someone’s not even willing to stake their job on it, I don’t see why I’d be concerned. (And until they do, we’re just speculating about what they may or may not even say.)
Is probably the best description of the state of engineering today
The ability for a model to improve itself over time without outside intervention. As yet, no models yet have this capability, and LLMs specifically what with how they are constructed, are not likely to ever have it. They are frozen in carbonite at the time of their training.
That's not to say they can't use skills to appear somewhat worldly, things like searching the internet and what have you, but if you ask a model to strictly stick to data in it's training, it will do that and it will be stuck in whatever time in which it was built.
Energy is conserved, gravity exists, and LLMs aren't intelligent.
We will never prove LLMs are intelligent, we'll only prove humans are not.
> can't use skills to appear somewhat worldly
Duh, humans don't generally make up things whole cloth either, they access other information, books, tools, sensors, etc.
But I guess if you ignore enough information you can reach any position. It's up to the audience to decide which one of us this applies to.
A model being able to train a bigger model with marginally better capabilities is not what the hype was about.
https://news.ycombinator.com/item?id=50355251
I think it's reasonable to assume that if it's not worth breaking the NDA, it's likely not an existential threat for the vast majority of humanity.
Now, does that mean that people who currently work upper-middle class knowledge work jobs aren't in for an exceedingly harsh economic reality check? Absolutely not.
But that's not really as serious as an existential threat to humanity.
On the one hand, you've potentially got an indeterminate number of future comfortable years based on a subjective assessment of how long you think it takes catastrophe to materialize, with the downside that you can't warn anyone and so no one will stop the coming catastrophe.
On the other hand you've got the immediate assurance of your current career being gone and high likelihood that you'll struggle financially as a result, with the upside that you can provide a warning to the public that they will almost certainly ignore and catastrophe comes anyway.
Given the risk/reward assessment of those two options, what do you pick?
Listen, I understand that, owing very likely to our differing backgrounds, you and I think differently.
Where threats are concerned, I'm a simple soldier and as such I see things simply.
I would raise the alarm without a second thought. I honestly don't know how anyone could reasonably believe they would have any other option?
Please understand, it's a job. I spot a threat to my unit, I report it up the chain of command.
It's not about whether or not there is a negative impact on me. It's not even about whether or not it serves a greater good; nor whether there is an ultimate benefit to my unit; nor whether I get brownie points when whatever you deem divine evaluates my soul in the hereafter.
It's my job to report. So I report.
I've never once considered the things you're talking about now when I was in any such situation. It was just a job.
Here's the thing, it kind of made me just automatically think a certain way. I know that if I see an existential threat, I'm reporting it. I believe anyone would. Not for benefit. Just because that's what you do. If you see a mass shooting in progress, you call the cops. I don't think there is a single person who thinks about negative impact or greater good. It's just automatic.
I was only saying that if the NDA would prevent them from reporting, then, to my mind, the threat is likely not existential. But, again, I do make room for the possibility that not everyone thinks like me. I just think it's very likely that people think exactly as I do when faced with that level of threat.
Maybe I'm wrong? That's fine.
The downside risk to yourself doesn't have to be existential for the upside to not be worth it.
Coming at this from a completely different angle, other people may see their duty to the law and the agreements they've signed as being more morally binding over their behaviors than their social obligation to raise the alarm. Not everyone uses the same formula to calculate which duty is most important.
Theoretically, retaliating against whistleblowers is seriously illegal stuff. But in practice these regulations are seldom enforced. And even if your (former) employer does not retaliate openly, there might still be "invisible" retaliation, or the risk of other large employers just blacklisting you by default.
[1] https://www.bbc.com/news/articles/cy42xgeq9vpo
May I ask? Why should I show faith and grace to the people actively involved in turning our society into a Big Brother program? People who work with the DoD to build war machines? Directly or indirectly involved, they either know what they’re doing or choose to look the other way, for money.
Please enlighten me, what is the faith and grace approach I should be stating?
1. Full disclosure. 2. Run for your life. 3. Live in exile.
2. sideload Signal
3. find a verified reporter on Bluesky, etc. with public Signal contact info
4. destroy phone thoroughly afterwards, not just screen but chips
If they want to do that free of consequences, tough luck.