I cannot fathom how you can intensely work on a problem that you genuinely believe has a "10 to 25%" probability of causing immense harm or even present an extinction-level risk to humanity. How can you truly believe this and be ok with it?
The early batch of Anthropic employees were mostly rationalist-adjacent AI safety folk that were almost uniformly claiming P_DOOM > .10 three years ago, so I believe them to be earnest.
It's very interesting to me that besides the other small safety labs that don't actually produce frontier models, Anthropic manages to keep such a good reputation within that subculture compared to OpenAI. Despite having as crazy internal politics as OpenAI, they have converged quite a bit from the original vision of safety first through Darwinistic pressures.
At least, it seems this way from the outside. I'm curious if the view from the inside is that different.
edit: to be clear, my reading as an outsider is that Anthropic is seen as relatively better in the AI safety community, but has definitely dropped in absolute reputation too. This recent thread and the references show some of that: https://www.lesswrong.com/posts/6j3kBHdowGLCeqobg/dear-god-p...
The Manhattan Project succeeded, so clearly some people are ok working on this sort of thing. And with similar justifications: if we don't build the potentially world-ending bomb, someone else might beat us to it.
That was openly a military project during wartime (and not being pitched as the solution to all problems at the same time so no cognitive dissonance required)
Both OpenAI (back in 2015) and Anthropic (much later, in 2021) were founded by people who could foresee the concept of AI x-risks and wanted to work on preventing it. For OpenAI's founding, the idea was that it's much safer if AGI is achieved by a nonprofit explicitly dedicated to humanity than if it's done by a profit-driven company. For Anthropic's, it was that OpenAI seems to be going insane and it'd be better if a more safety-conscious company was competing. I'd say in hindsight, the former motivation was reasonable and the latter one was... probably bad, but not obviously so - it doesn't seem even in hindsight that it was inevitable that Anthropic's existence would only drive competition and cause them both to race to AGI. That said, even at founding time, there was immediately a lot of selection - quite a few people who cared about AI safety simply wouldn't agree to work at a capabilities company, even given an elaborate argument why this is a good idea, so those were selected against.
And then, of course, many years passed, OpenAI was stolen by Sam Altman and turned into a for-profit, multiple researchers left OpenAI realising that they're only helping build misaligned AI faster, multiple researchers left Anthropic realising that they, too, are only helping build misaligned AI faster, and here we are in 2026.
There's lots of AI researchers these days, so you can select for conformance very heavily and still have a fully-staffed company. The people who are still at these companies are outliers in various ways. They either manage to dismiss the importance AI risks (for example, by adopting some sort of belief in the vein of "there's no use worrying about AI wiping out humanity - it'd just be a successor species, like we were to apes, which is good"*), or they still think that leaving will not make things go better (Dario Amodei is pretty clearly in this camp, and has always been), or they weight the risk of extinction against the potential benefit of an post-singularity utopia and consider it a good bet (not realizing that the alternative isn't giving up on a post-singularity utopia, but getting it a few decades later and without the risk), or they just manage not to think of the contradictions, which humans are of course very good at.
Excellent point! Considering the history of these companies definitely makes it clearer how you could reach this point over a few years similar to the boiling frog story. A slow moral decrepitude and moving goalpoasts in the name of "national security" and "humanity's safety" over a backdrop of ultra-utilitarian rationalist thinking ("if it's not us, it's them")
There must have been thousands of people who built or did work related to nuclear weapons in the Cold War, people are just really good at rationalising away probabilistic dangers with unclear consequences, especially when there’s a corresponding upside. The same is true for climate change, drug addiction, health problems etc.
I also think some of them have become delusional and convinced themselves that AI is going to create some kind of transhumanist utopia, even Dario Armodei leans in this direction from time to time. I imagine the people in these labs spend much of their day talking to sycophantic AI models that will encourage their delusional ideas.
Do you genuinely believe people always act logically? I don't know a single person that does. Don't underestimate the human capability to hold several completely contradictory beliefs with zero self-reflection. There's physicists who believe in god, for fuck's sake.
> How can you truly believe this and be ok with it?
Are you saying they don't truly believe this, or that they aren't OK with it?
My personal opinion, this doom-mongering about LLMs, is part of marketing campaign to exaggerate LLM capabilities. (Play being: Sociopaths with money will buy or invest if they believe LLMs are powerful things)
If you are not familiar with Camus' work, I highly recommend reading the essay collection Nuptials (Nuptials at Tipasa, The Wind at Djémila, Summer in Algiers and The Desert). It was written in 1936, when Camus was in his early twenties; several years before he articulated his "philosophie de l'absurde" in The Myth of Sisyphus and The Stranger.
Camus himself described Noces as belonging to a pre-absurdist "pagan" phase. It describes, in beautifully poetic french, the intense, sensual love of physical existence and the here-and-now that gives the absurd its stakes.
A quote from Nuptials at Tipasa:
In the spring, Tipasa is inhabited by gods and the gods speak in the sun and the
scent of absinthe leaves, in the silver armor of the sea, in the raw blue sky,
the flower-covered ruins, and the great bubbles of light among the heaps of
stone. At certain hours of the day the countryside is black with sunlight. The
eyes try in vain to perceive anything but drops of light and colors trembling on
the lashes.
[...]
I love this life with abandon and wish to speak of it boldly: it makes me proud of my human condition. Yet people have often told me: there’s nothing to be proud of. Yes, there is: this sun, this sea, my heart leaping with youth, the salt taste of my body and this vast landscape in which tenderness and glory merge in blue and yellow. It is to conquer this that I need my strength and my resources. Everything here leaves me intact, I surrender nothing of myself, and don no mask: learning patiently and arduously how to live is enough for me, well worth all their arts of living.
For me, it’s the following sentence that gives the meaning to that one.
The countryside is literally so beautiful that the beauty cannot (be) beheld [sic]. That is, the sunlight hitting that bit of the world at that bob of time is so intense that the viewing is indistinguishable in its ordinary pieces because of the sheer pleasure. The view is just a singular whole all at once, because the thick and luminous real is everywhere and dripping off of every thing. The abundance drowns out the particulars.
This is great. My enhancement was to create a .claude-profile file which specifies the account alias. "claude" can be aliased to read the file, set the correct env var, and launch claude.
Another benefit was if no .claude-profile file exists, claude refuses to launch.
Depends on your point of view. Has the web been a net positive overall? If your criteria is "instant access to any information" or "catalyze more technological progress" then positive, if your criteria is "young people's mental health" or "how easy it is to control the public's perception about a certain topic" then negative in my opinion. It's all relative.
Short form videos (tik tok, YouTube shorts, instagram reels)
Coming soon:
Deep sea mining
Some of these are not great technologies, you might say. Well if they're impactful enough to turn everyone I love into a phone zombie and make the political landscape violent and insane then that's enough.
30 years ago was the 90’s, when just about everything peaked.
The things I’d be truly missing from today 30 years ago, would be the advances in medicine really and the fact that we don’t have the ozone layer threat.
Most other things have peaked back then. Rising kids was definitely easier, and healthier.
All frontier models have been trained without any regards for IP protection laws. I don't see how anyone can argue in good faith that distillation is not fair game and does not ultimately "benefit humanity™"
I get what you're saying and two wrongs do not make a right but the irony, and why people are even talking about this, is that the thing being distilled clearly, and knowingly, violated copyright & terms across the entire internet.
For me, I don't really care about the theft aspect but when people are claiming that these open models are better value or going to overtake anthropic/openai models, the implication that the open models are training of distilled data means all the "progress" they are making is just mimiced from the closed models.
It's a bit interesting how the open models are able to keep pace with the closed models except whole maintaining a steady following time.
It's important to note that even if these open models are distilled, they are showing genuine improvements in their architecture, which enables inference costs to be several factors below what equivalent closed models have.
The interesting question is: will Anthropic release a Fable like model with an architecture similar to Kimi, and get the inference cost gains? They should surely beat Kimi because they can internally distill as much as they want.
Distillation is absolutely not the reason they're good. It's not necessarily even done on a more capable model. It can even be done on itself and still bring improvement, or on a weaker model as well (see GLM and Gemini, which is definitely true because it repeats Deepmind's injections).
It does, it's just not the reason. Most of the work is done before that point, and as I said z.ai used a weaker model. Besides, this is all strictly one-sided, as nobody knows how much Anthropic and OpenAI borrowed from Chinese labs' open research and weights (and they innovated a lot, to put it mildly, starting with first reasoning models worth talking about long before OAI did the same). Chinese labs are also severely restricted on hardware.
This entire story makes certain American AI shops look cartoonishly evil and Chinese ones relatively sane. Not only they want to grab without giving anything back, they also want to sabotage everyone else's AI research and do plenty of terrible things like media manipulation on the global scale and getting in bed with the government. This can't possibly end well, for the Americans in the first place.
Training a model is very expensive and creates something no individual rights-holder could. Distilling a model copies this value add and captures it without bearing the cost that created it.
They only had to pay for storing the books on a server for later possible use. They did not have to pay anything for the training which was declared fair use.
This actually undermines the argument that distilling is harmless because its founded on the idea that Anthropic did the same thing and didn’t have any repercussions.
I understand you think Anthropic should have paid for the information it trained the models on. But im talking about all the costs to build a model. Do you think Anthropic didnt spend money to build these models? Did you not know that its actually very expensive?
Moonshot pay for access at the price point Anthropic sees fit, potentially more, since they likely had to jump through multiple hoops?
It's hard to feel sorry about the breach of their ToS, which ultimately is all Anthropic can argue, when they are constantly being sued by countless IP owners
Could you give an example of the value that only training a model can create but none of the rights-holder could? I feel like if you got a direct, instant communication channel to any of the rights-holder that created the content in the training set of those models, you'd get more value than what the LLM could ever give you on any specific subject.
the outputs of the model have no property protections, and training a model on the outputs of another model does the exact same thing - its expensive and creates new value over what was in the input - a set of documents.
Exactly. Imagine what would be possible after a billion year of technological evolution, heck even just 100'000 years. We already know that space time metric engineering is theoretically possible within our current understanding of physics, we don't have either the technology or access to energy density necessary to do it. And that's only within our limited understanding of how the universe works.
If you have a billion year long civilization, it would probably be easier to genetically/medically engineer humans to live forever, so that a 75,000 year space voyage doesn’t seem so long. Easier than near-lightspeed space travel at any rate.
The general idea is to deliberately shape the geometry of spacetime to allow for effects such as faster-than-light travel (relative to another region of spacetime).
Einstein's field equations link geometry to matter and energy:
G_μν = 8πG/c⁴ · T_μν
Left side describes how spacetime curves (the metric tensor). Right side represents what's causing the curvature (the stress-energy tensor): mass, energy, momentum. The usual approach is: you specify T_μν (what matter is where) and solve for the resulting curvature. Metric engineering takes the backward approach. You pick the geometry you want first: e.g a bubble that carries its contents faster than light relative to the outside universe. Then you work out what T_μν would have to be to produce that geometry: what distribution of energy and stress the equations demand.
The classic example of spacetime metric engineering is the Alcubierre drive.
The limitation with this is that the required T_μν has negative energy density in some regions (the bubble walls) which violate some assumptions about how "reasonable" matter behaves. But Quantum Field theory allows such local regions of negative densities and we have created such regions in labs experimentally already (Casimir effect). We "just" don't know how to sustain the effect at bigger (atomic) distances.
Neo4j is a great DB but their license price is egregious for enterprise customers. A few years ago I was involved in negotiating a contract for a small/medium size kubernetes deployment (think around 25 cores) and the annual price was more than the salary of a senior SWE full-time equivalent. See this page for an idea of their prices in 2018: https://blog.igovsol.com/2018/01/10/Neo4j-Commercial-Prices....
Also embedding Neo4j is not possible, that seems to be the killer feature for YouTrackDB, they even shade dependencies so it’s like a no deps Java library for your application.
> small/medium size kubernetes deployment (think around 25 cores
That's ~1 machine. 1 SWE for a database isn't egregious, databases provide huge value, but for that little performance, that's crazy.
I can only assume as core count has blown up over the last 10 years, the pricing has somewhat diminished, but still, I'd be expecting a heck of a lot more capacity for 1 SWE.
I can tell you from quotes that I've seen, that compared to the 2018 prices listed on the page, end-of-2022 pricing was slightly up per core.
We were already well on our way of dumping Neo4J due to performance/operational/architectural reasons at the point, but seeing the quote solidified it.
Not just that, if a database company has both a community edition and enterprise then it’s likely the enterprise will get many new features that the community edition will never get.
Even more so considering that most, if not all countries already have age-verification infrastructure in-place for selling alcohol and tobacco (and other goods). Why not piggy-back on this? Go to a kiosk/shop selling controlled goods, pay a small fee to the vendor to get a one-use verification code that attests your age.
Making someone get in the car and go somewhere to be able to use a website is a rough ask.
Just a thought, maybe I could get a one-time code from my secretary of state website that provides only the age bracket information to the service provider’s age verification broker.
- Don’t need to go anywhere.
- Doesn’t provide age information to the service provider beyond “yes/no, allowed”.
- Doesn’t tell the agency who knows your age what service you’re trying to use.
- Broker doesn’t know any actual identity or service details, and can ideally be ZDR
This is literally the purpose of most of the European EUDI wallets, but automated. You create an account with whatever part of government handles ID, you get a "wallet", and then services can ask you to share minimised data from the wallet. This can be implemented in such a way that the government doesn't know who requested and the requester doesn't learn anything about you other than the minimum covering set of attributes required.
Love it. We've been playing around with other ideas that are not web related. As the tech improves, less data will be needed to infer age. Strange now, but setting up a microphone at a kiosk or providing a widget-less API for broader use-cases, is not a stretch for piggy-backing. Thank you for the comment.
> We are not seeking job displacement. We are working to prevent or minimize it. Some amount of displacement, though we cannot say how much, may be an intrinsic consequence of the technology, and our responsibility is to prepare for it and respond to it. That is what this framework is for.
> Whatever happens, we are on the side of people. We are trying to solve these problems. We take no satisfaction in contributing to them, and we are not working to make them more likely.
The cognitive dissonance/doublespeak/hypocrisy (pick one) is absolutely insane.
They are concurrently:
1. creating and marketing products that are explicitly trying to automate, if not entire professions, at least big parts of them
2. edicting grand policy plans to limit the impact of massive job displacement that their products might cause
3. directly funding and coordinating missionary-type activities ("it's for a greater cause") to evangelize and propagate said products in areas of the economy that are usually underfunded and where job security is already quite bad (non-profits, NGOs)
So what do you want them to do, then? Because I don't like it either, but they're right about their core tenet, which is that if they don't do it, someone else is going to. Even if we ignore the obvious other player in OpenAI, there's now dozens of labs in a handful of countries, the cat is thoroughly out of the bag.
Is it annoying? Yeah. Is it preachy and hypocritical? Yeah. And I've yet to hear anyone suggest something better.
> And I've yet to hear anyone suggest something better.
While I reject their basic premise and think it is more marketing garbage, the answer sure as shit isn’t to try to monopolize the technology so that Anthropic can extract maximum rents from it.
Right. Five years ago the idea was you needed the capitalism to fund compute/training. Now, its clear that open weight models could be a public good. If my PC takes my job I'm still empowered and can still capture that value, so if Anthropic were pursing that end it wouldn't seem so hypocritical.
Continue building useful models, tooling and products, market them to the people who could benefit and be realistic about their strength and weaknesses. Drop the messaging about job displacement, "transformative/dangerous AI", "significant disruptions" and "unprecedented abundance". They are the ones pushing and funding this narrative; just look at the phrasing in this announcement. And it's percolating down everywhere in the general public leading to AI psychosis.
They are proposing solutions to a hypothetical problem they are actively trying to manifest into reality to get more capital and funding.
That's what hypocritical about it. Not the development of the technology nor the effort to lobby for more regulations and policies.
The general discourse around AI could be much more sensible and pragmatic and lead to a more balanced, healthy rollout of the technology in society. But of course this means forfeiting at least part of the massive injection of capital in those companies and the ecosystem in the short-term.
I don't understand this at a high level. Anthropic is going to go public, which means they are liable to act in the best interest of shareholders, so they cannot leave money on the table, much less do selfless handouts.
Yet at the same time, they are a public benefit corporation, and I'm not quite sure what that means? Is that a separate legal entity from the for-profit arm?
Them being a public benefit cooperation means that Anthropic is legally required to consider both shareholder interests and the broader public good. At least that is how they explain it: https://www.anthropic.com/news/the-long-term-benefit-trust
There is also a separate legal entity (Anthropic Long-Term Benefit Trust) that owns some shares and has special powers, but shareholders can change that so I wouldn't trust the trust too much.
Hey, think critically about where you are on the internet. You're on a message board run by YCombinator, who's stated goal is to teach people to run startups. Startups are inherently disruptive. When one business disrupts another, people lose their jobs.
Companies going out of business, either because of disruption, or because they ran themselves poorly, or other reasons, is part of the normal business cycle. Otherwise, we'd end up doing things like making digital cameras illegal because the people who worked in film labs lost their jobs. (Which is absurd!)
---
I don't see any cognitive dissonance in what you quote. Some people will lose jobs to AI. Anthropic wants to train people who lose jobs to AI.
I see your point, but in this case we are not talking about disruption at the scale of a product category, vertical, or even an entire industry. AI companies are trying to disrupt entire sectors of the economy at the same time: knowledge work/white collar jobs, creative work (design, photo, video, ...), medical professions, etc.
They are recognizing themselves in their economic policy framework that the lowest level of unemployment potentially caused would be 5% (they also mention 10% and "unprecedented levels of unemployment").
I don't think there is a precedent for this claim. It's hard to take the "we're a force for good for humanity" message seriously in this context.
AI companies are trying to disrupt entire sectors of the economy at the same time
Talk to news companies, which whether newspapers, magazines, radio or TV, are basically either literally gone, or a shadow of their former selves. The best way to compare this, as you say is to even larger constructs, to entire sectors of the economy.
Because you can harken this to desktop computers, or to the Internet. Sector after sector of people replaced, industries gone. Just the sheer amount of newsprint, recycled or not, which is now no longer manufactured affected an enormous amount of jobs. The entire forest industry, to sawmill, to shipping + printing and all the parts and maintenance and delivery and even the newstands which are basically now gone. Good for the environment or not, that's massive change.
I agree AI will do the very same. But it's not even really happening that much faster. We're 3 years in. It will take 10 to 20 years for it to play out.
But... back in 2008, would you have said the Internet was "bad" for humanity? No, you wouldn't. Back then, it was all about connecting people, it was about empowering people in totalitarian regimes, so they could connect and speak out. It was about old controls slipping, about people being able to speak directly to one another. Back then, 99.999% of people thought "awesome!".
And when I talk to the average Joe about AI today? I hear the same thing. Awesome!
The Internet has put countless people out of work. So has the computer. So has electricity. And machines. People have always complained. But if you're going to label AI companies as "They took my job!", then you'd better do the same for all those other industries. Otherwise?
I was not criticizing the technology in itself but the messaging and marketing around it. I agree with most of your points except perhaps the general sentiment outside of tech about AI. Most people I've talked to admire the technology but are genuinely scared or angry about the second-order effect of it on society. And I think, in part, it's because of the messaging, marketing and general psychosis about the technology that those companies are generating.
Sure I'll do the same for those other industries. It took generations and many deaths after each industrial revolution to address the harms caused by the technology. People's complaints about economic disruption have been proven to be legitimate and require fighting. This will spill into violence if these criticisms are not taken seriously.
And yet unless you're a historian, most people are completely unaware of any of the violence which occurred as a result of prior disruption. Most aren't even aware there were protests, upset in the past, except perhaps due to a song about a miner, John Henry.
These prior points of 'fighting' and 'violence' are so inconsequential, that they don't even impinge upon the historical narrative of most nations. People know about war, civil war, even the hippy protests half a century ago.
And in the end? In every single case? Nothing happened to realistically slow that change. It happened regardless.
One of the issues with such change, is that we live in a world with national borders. This leaves individuals in nations with two choices. Improve efficiency (by replacing humans with more efficient methods), or alternative? Companies with more efficient methods, in other countries, will drive your local companies out of business.
Which means you lose any tax revenue, resource extraction revenue, "head office" revenue, and on and on. And of course, whatever humans are level after the reorg, "management", well those jobs go to that other country too.
Literally if you win in terms of preventing change, you lose.
In the context of all of this, legitimate or not isn't really relevant. And fighting or not is in that same category. Nothing will stop the changes coming. Nothing.
If we were in the same room, you would hear a typically exceptionally optimistic person, speaking in a very sad voice.
Also, they keep talking about the US like its the only country in the world. They where fine with Claude being used murder and spy as long as it was done to others. And again they are only talking about the US work force. Howany engineers in India will loose their jobs when the ai becomes cheaper? What about other Asian countries with low wage jobs that will be replaced?
> They are recognizing themselves in their economic policy framework that the lowest level of unemployment potentially caused would be 5% (they also mention 10% and "unprecedented levels of unemployment").
Every tech boom I've seen in my life includes people letting their imagination running away and greatly overestimating how adoption will take place. Often adoption is slower, but much more pervasive as younger generations use it when they become adults.
Translation: I believe AI's disruption will be much slower than a lot of people claim it will be; and the change will be (cough) "slow" enough that most people will be able to retire or change their careers before they are materially impacted.
They are not, the disruption itself becoming the praiseworthy goal was not there initially. Initially, the rhetoric was about creating new things, building companies and booming economy.
It changed into disruption only later. At some point, disrupting became the praised thing, even if you was in the loss the whole time and did not really produced anything better.
> Initially, the rhetoric was about creating new things, building companies and booming economy.
The "new things" always disrupt something else. IE, a mechanized excavator disrupts people who made their living digging ditches.
BTW, were you around Silicon Valley in the 2005-2015 decade? The people making AI were easy to find and were very clearly "about creating new things, building companies and booming economy".
> economy that are usually underfunded and where job security is already quite bad (non-profits, NGOs)
If it works, using AI to cut jobs in the non-profit sector would be a good thing. Those jobs are overhead (necessary overhead, but overhead nonetheless) that reduces how much donor money actually makes it to the people it’s supposed to help. If AI can eliminate those jobs, that means a larger share of money that’s donated will actually get used for providing services.
> edicting grand policy plans to limit the impact of massive job displacement that their products might cause
Your own job, assuming you're in the software industry, is to automate and eliminate people's jobs. How many jobs do you think it has eliminated? Do you feel you are responsible for your products? Is it ethical to do your job without thinking about and discussing its impacts?
My take: being also in the industry, software has created way more jobs, new businesses, new fields than the jobs it has eliminated. Nobody knows how AI will turn out, it being still at the beginning, although it will certainly have big impact and job displacement will be substantial. The hope is that it will be net plus for the society. At least Anthropic is talking about it. I never heard of, for example DeepSeek's position about the impacts of their products.
> directly funding and coordinating missionary-type activities ("it's for a greater cause") to evangelize and propagate said products in areas of the economy that are usually underfunded and where job security is already quite bad (non-profits, NGOs)
If you've volunteered for non-profits, you'd find that many of them are underfunded AND understaffed. Removing burdens on any part of their work, especially areas that aren't directly related to their core services, is hugely beneficial. It's easy to criticize from behind the keyboard.
Unfortunately opinions like yours scratch an itch of the HN crowd. Regardless of objectivity.
> Your own job, assuming you're in the software industry, is to automate and eliminate people's jobs. How many jobs do you think it has eliminated?
Looking at my career, literally none. For real.
> Is it ethical to do your job without thinking about and discussing its impacts?
I am not saving the planet nor people. I am not social worker, doctor nor anyone like that. But, I am not harming them either.
> At least Anthropic is talking about it. I never heard of, for example DeepSeek's position about the impacts of their products.
No, DeepSeek is not trying to create FOMO and fear based sales. DeepSeek is not pushing insulting ads on me and scary ads on my CEO. I dont even think they are all that much more ethical as a business necessarily, but oh my god, they managed to avoid the most off-putting propaganda in places where I see.
> Your own job, assuming you're in the software industry, is to automate and eliminate people's jobs. How many jobs do you think it has eliminated? Do you feel you are responsible for your products?
This is a common trope from the LLM crowd. Places I have worked for as a software engineer have created more jobs (gig economy) or improved the human condition (edtech in emerging markets) or helped people in refugee camps in Kenya stay connected.
Even in questionable companies, I focus on work that makes sure the technology is accessible to any and everyone. I became a programmer because I thought I could help make the world a better place.
> The hope is that it will be net plus for the society.
“We are not seeking virtual torture caused by Roko’s Basilisk. We are working to prevent or minimize virtual torture caused by it. Some amount of virtual torture, though we cannot say how much, may be an intrinsic consequence of the technology, and our responsibility is to prepare for it and respond to it. That is what this framework is for.”
How is it difficult to make the connection between a thought experiment about a world-eating AI monster, Anthropic marketing their AI as a world-eating monster, and Anthropic accelerating frontier development of their world-eating monster as hard as possible possibly being a little tiny bit icky and cynical when they tell us they’re totally trying to save us from it and stuff, bro?
I think I'd place less emphasis on the "creating" part and more emphasis on the "marketing" part. If Dario could shut up about how jobs are outdated, that would be nice. His quotes simply give more ammunition to incompetent CEOs trying to cut costs.
It also goes without saying that a certain segment of jobs are simply bullshit jobs that are prone to automation anyways. But without those jobs, you're also cutting off a segment of the population from the economy.
It doesn't matter if they can or cannot. Every marketing copy in Anthropic is about replacing workers. Even their job listings state "This position may be replaced by AI within a year.".
reply