AI as a tech is fine. But disliking it and the social/economic effects around it is fine too, people should be allowed to feel however they want to feel about certain techs and situations.
To recommend people to suck it up is not the answer I wish in the society I want to live in.
There's a normative argument in the parent that's reasonable to engage and rebut, but there's also a positive component that's less easy to take issue with. It really isn't going anywhere, no matter what world you want to live in. People were upset about databases in the 1980s (some still are).
What makes you say that AI is not going anywhere? I hear this overwhelmingly, "AI is here to stay", as if y'all are so caught up in the movement that you've started taking that conclusion as being the axiom. TBH, it feels like a religion.
Short of societal collapse, there's no way the technology is going to go away or fade out of existence (unless it's replaced by something even better), that's just not how technological progress works. It's useful, probably in ways we haven't even thought of yet.
I don't get it, why would operating a datacenter needs massive amount of high skilled blue-collar labor. Datacenters are resource hungry. With so much automation in place I don't think there would be a need for large pool of labor.
You seem to suppose the building of those datacenters - even the power plants behind them - won't soon be automated. Almost as if robotics isn't happening.
They stopped being used as primary weapons because better ones were found - mostly firearms - not because people got bored of it; or reverted to some earlier methods of warfare.
Yes, there is the general class of technologies (warfare, computing ...) and there are particular instances of those for a given time and space and evolve as the landscape changes.
The technology of warfare evolved to better mechanisms, perhaps same with computing.
Bows and arrows are still widely used for hunting all over the world. I was able do freelance work on a relatively low income because of access to ~150lbs of deer meat that came from multiple bow-hunted deer.
This is an interesting question that I haven't thought about, thanks.
What we currently have is a simulacrum of thought - albeit a good one.
Any technology is useful only in the sense that it helps us with solving the problems we are dealing with in that time. When we face issues that a pseudo-thought is not useful in tackling or worse is one of the causes - this will recede in the background.
Beyond that, the implicit assumption in the question is that thinking is the highest form of activity that is useful to us.
I don't know how my thoughts arise but thinking happens when I engage with them. I think what we look for is meaning in our lives and thinking helps us generate/achieve one, whether real or illusory.
I don't know about you, but I can buy bows and arrows at hundreds of sporting goods stores in my local area alone, and I even know of 2 local blacksmith shops that sell swords.
Castles still exist as well, you just aren't invited to them (which was true for us peasants back in the day, too). Trump is still trying to get one built under the ruins of the East Wing, in fact.
But are these actually completely different technologies, and if so, where is the dividing line? Firearms certainly have not decreased in significance, and they're the modern version of a bow, which is simply 2 iterations later in propulsion methods: tensioned string -> high-tension cable -> high-pressure gas.
Are LLMs really going to fall off in significance, or will it just be the nth newest incarnation of LLMs?
The function of what an LLM does (generative language) is what people seem to take issue with, but the function is here to stay, even if the next iteration has a different name or method.
The difference is in what other enabling technologies do you need to achieve it. Advanced technologies sit on a pyramid. One can build a bow and an arrow from sticks, string and rock. For reliable firearm we need chemistry and advanced metallurgy.
My view wasn't whether generative language is here to stay or not but rather will it continue to be a significant thing or not.
The technology involved in Juicero (or Pets.com, or many others) didn't go away. We could rebuild them any time we wanted to. Those things went away because they weren't able to make enough money to be an ongoing business.
Will AI? That is at least an open question at this point. (I mean, in fairness, Amazon's was an open question for many years too.)
The tech isn't going anywhere. Is there a path to a sustainable business model that uses that tech?
You may have an answer to that question. Can you prove it to someone who doesn't already agree with your answer?
Juicero wasn't useful therefore it went away. Generative AI is useful therefore it won't go away, just like how fire is kMy old yet it's still here to stay.
Juicero had customers that thought it was useful. Both "smoothie juicers" and "smoothie subscriptions" were big product categories equally before and after Juicero tried to unite them with a technological middleman. Kuerig, the Juicero of coffee, remains quite active and profitable (and wasteful).
There are likely _many_ paths to sustainable business models based on AI tech, that will come to fruition over the next decades. However whether they might not be as profitable as OpenAI and Anthropic are gambling on, is more uncertain.
In the same way that any technology could just magically disappear, sure.
But I hear everyday, non-IT-sector people talking constantly about how they're using it, and that means there's a demand for it, and someone is going to supply it. I think a lot of anti-AI people think it's still equivalent to the PDA, and don't realize it's a smartphone already.
The other side is that "AI" is of course very very broad and isn't new, and e.g. medical vision models are making advancements that are having huge impacts on patient care already, especially around early cancer detection. Those aren't going away (and shouldn't), so there's still going to be a demand for the underlying technology and infrastructure to support it, even if LLMs stop being spammed everywhere.
The other thing which people seem not to understand is that you don't need a whole datacenter to RUN individual LLMs, you need it to train them, or to run them at scale for thousands of customers. A lot of the upper-mid-tier models that exist now can be run on a single (beefy) 4U server in your closet if you've got the GPUs to put in it. And people are running e.g. Deepseek V4 Pro FP4 locally. If you've got an actual server room, like at a university, you can run the full, un-quantized versions with ~2-4 servers.
Technology that is living in peoples' homes and businesses already is not going to just disappear. It's a lot less centralized than the market prevalence of OpenAI and Anthropic would lead you to believe.
I think this disconnect is based on the ambiguity in the term "AI".
"AI" as tech - the models, how to train them, etc. Isn't going to go anywhere short of a Library-of-Alexandria-type catastrophe. We know how to do it and it's useful, so why would we forget?
However, "AI" as the thing that is enveloping our culture - the slop everywhere, the mandates to use it at work regardless of its usefulness, the constant talk about it being the future, the machine-dominated future that's been promised/threatened by the heads of the labs - we do still have a chance to put that onto the scrapheap.
The old gods lost their power because they had been conquered by the Spirit of Christ.
It's a danger to beware of, that people might start to listen to LLMs like they have the gospel truth on morality and spiritual matters. But they're just an image of a man given breath and speech, without true wisdom. Man took sand and copper and formed the simulation of a mind. If we then bow down to it and obey it, we're fools.
It makes tremendous sense - when understand as reflexive straw-clutching and wish-thinking aimed at reducing the frequency of the poster's nightmares and reducing their diaper expense.
Sweden had from 1973-1998 a law that made it illegal to have a computer database of personal information without getting approval from the government (in 1982 it was opened up so that approval was only needed for "sensitive" information).
Looking back getting rid of that may have been a mistake.
AI is here to stay, I don't want it anywhere near the art, literature, and music I enjoy, not least because part of the enjoyment comes from the knowledge it had a very human creator. That should be perfectly achievable.
People forget that a lot; my father came home end 70s explaining his life was over because databases, mid 80s because code could now be synthesised from models (with 'AI') that domain experts write; the latter went on a bit in different forms until now where it is becoming reality for things that were not very hard before anyway or in the hands of people who use it as one of their tools (antirez comes to mind), not as 'english programming'. The absolute crap (ads, tracking, no responsibility because computer says no etc) my generation built is, in my eyes, not really positive without something to counter it. Many positive things are there, but many things 'we' started and made normal must be ring-fenced and controlled as they are negative to an absolute sometimes. The current AI is hard to see; I am building things with it I could have never built on my own (and I have been programming since the 70s) as programmer, tech lead or cto, 1000s of projects over the decades, some tiny, some huge. I could build complex things but they took time, now they take time but only a fraction. But what I see most people building is absolute slop; it has no function outside trying to sell something that has no value in a time you still can if (and only if) you can do a little dance on tiktok for an audience. I will keep on happily hacking anyway until I die.
> But what I see most people building is absolute slop; it has no function outside trying to sell something that has no value in a time you still can if (and only if) you can do a little dance on tiktok for an audience
people are doing this inside companies as well; stupid animations generated by ai posted in slack, powerpoint slides with super detailed drawings and images of ai generated "user experiences" where a couple of text boxes would have been fine, etc etc
The widespread criticism isn't based on "the tech not going away eventually". All of that is a red herring, a response to something nobody (mainstream) is talking about.
Crypto bros said the same thing about NFT’s and ICO’s and whatever other nonsense they were pushing. And to some extent, they were right, I guess, in that these things still exist. But they’re practically irrelevant.
There's plenty of things that are ubiquitous but not well-liked, so I don't see how "it's not going away, get over it" works as an argument. Many people won't be getting over it. Traffic jams are here to stay but I'm never delighted to be in one.
Outside the tech bubble, a significant proportion of the population is using AI, but in all surveys, it's hugely disliked. It's probably due to social anxieties that in big part trace back to how AI tech companies do marketing. If you have billboards that say "don't hire humans" and Gates and Altman talking about how most jobs are going away, what do you expect? People are not gonna be optimistic even if they secretly enjoy asking ChatGPT for relationship advice.
I think that AI is less analogous to "traffic jams" and more analogous to "wheel-based transportation". It's an entire category, not a specific problem. The traffic jam is more analogous to excessive energy consumption or workforce disruption.
Many people seem unable to separate the wheat from the chaff when it comes to AI.
> Many people seem unable to separate the wheat from the chaff when it comes to AI.
its because "ai" means whatever the marketing dept decides it is... we've had all kinds of 'ai' in the past its just they enabled features based on some intelligent logic not just llms.
right now llms are being shoved into everything whether its a good fit or not; so it makes for additional noise...
Some people say that we cannot solve catastrophic climate change. And then some other people claim that those are anti-solving the problem. Indeed the climate change problem is massive and it is incredibly, incredibly difficult to solve given the kind of world that we have engineered for ourselves. By contrast it wouldn’t be a problem at all to magically wipe the wonders of AI since that only happened three years ago, or last month, or last December, or whatever the current inflection point is or was deemed to be.
So I don’t really buy the inevitability of technological progress in a world where infinite progress and growth have turned out to be false. Especially with the strange dichotomy of this being so apparently obvious, as commonly stated, juxtaposed with the horde of people that point this out to us on the daily.
Tangentially, I expect both this Pandora’s Box narrative to continue and narratives about how the good times for commoners are over and they need to learn some real life skills like foraging for their own food. Just as a sort of emergent narrative development.
Germany shut down all their nuclear power plants. The Amish don't use tractors or whatever. I guess we could have AI free countries. The last bastions of true humanity or something.
I would not recommend that people "suck it up", but I think people have to come to terms with the fact that AI is a legitimate technology that is going to transform the way people live and work. That is just a fact of life, as surely true about AI as it was true about the internet, or smart phones, or cars, or radio, or the train.
You can close your eyes and pretend that it is not coming, or you can organize politically to mitigate the damage it is going to do while harnessing the benefits of it. Because it absolutely _is_ going to harm a lot of individuals, even if the best case scenario of benefiting humanity as a whole comes to pass.
There is no possible universe where AI is banned, or it just fails and goes away as a technology. None. People have to just accept that and focus on realistic ways to regulate it and tax it, instead.
This is the central problem with the dismissals of the tech's capability. Public discourse needs to shift to planning for the economic impact in particular, but the kind of High Brazilism from the naysayers who insist it's a proof of psychosis to even mention AI's potential, makes the inertia in policymakers much easier for them to maintain. Waiting for the financial effects to arrive and then improvising policy is the stupidest way of handling an upheaval on this scale - even if the precise form of those shocks can't be anticipated.
There's a voluminous amount of code and documentation on how to build and run LLMs. You can build your own chatgpt literally in a weekend and run it on a home server, based on publicly available models.
If OpenAI and Anthropic literally evaporated overnight, there would still be Chinese labs training and releasing new models.
I'm sorry - but you're not going to ban AI no more than you can ban the transistor. You could limit & limit the potential of who uses it - but historically that seems to benefit the few rather than the many.
Banning doesn't mean the thing you ban disappears, you're stretching the definition quite a lot. The limit you're talking about is what a ban is. I would personally support a complete ban on AI technology impersonating humans. At no point would I expect that would completely get rid of the phenomenon. But it should be seen as extremely cringe and face meaningful push back, given how antisocial and antihuman that is.
I don't want to live in a society where privacy is a second class citizen, yet the prevailing sentiment seems to be "suck it up, we have to protect the children."
There are plenty of things not to like about society. What's funny about AI is that the inequality it brings is proportionally affecting white people, and thats got a lot of people discovering with great consternation that the world isn't fair.
>people should be allowed to feel however they want to feel about certain techs and situations.
All the white collar workers whining about AI didn't give a damn about the tens of millions of factory workers who lost their jobs to automation. Society doesn't owe them any more sympathy than they gave to the workers whose jobs they automated away.
> To recommend people to suck it up is not the answer I wish in the society I want to live in.
Funny that AI is the breaking point here instead of things that actually matter. It's perfectly fine to tell people to suck it up on a huge number of more important issues.
> It's perfectly fine to tell people to suck it up on a huge number of more important issues.
No, it isn't. If you think it's "perfectly fine" to dismiss people's legitimate concerns and complaints by telling them to "suck it up", the problem is that you're an asshole, not that AI is unimportant or whatever it is you're trying to imply.
>to dismiss people's legitimate concerns and complaints by telling them to "suck it up", the problem is that you're an asshole,
Those same people were callously telling factory workers who lost their job to automation and outsourcing to "learn to code"; they don't deserve any sympathy. Assholes are the hypocrites who are fine automating other people's jobs away but not their own.
Ignoring your rudeness, the word "legitimate" is doing a lot of heavy lifting here. It would take me one second to present you with an issue that concerns me, that will make you say "suck it up" because you don't consider it a legit issue, and I would end up being the asshole in the exchange.
> It would take me one second to present you with an issue that concerns me
Does this hypothetical issue concern you AND the rest of society as a whole as well, or just you? Because there is a big difference between the two cases.
How many people do you require for it to be a legitmate concern? I can show you millions but you will disregard them anyway, because they all have wrong opinions.
I don't know where to draw the line entirely, it probably depends on a mixture of absolute numbers, relative numbers to the population involved in the discussion and the issue discussed. But it's clear that in this AI discussion case, telling "suck it up" to all the people complaining it's rude.
> Funny that AI is the breaking point here instead of things that actually matter
Nah, it's just one with high relevance to a tech audience. We say similar things around here re ubiquitous surveillance tech, internet censorship by governments / payment processors, the effects of social media...
Things that actually matter have been teetering on the edge because of the simple fact that labor has been needed to make money and money is power. If AI takes away the last leverage of labor, then things that actually matter will collapse entirely.
Could you give an example of an issue we tell people to suck it up about? I'd like to think a reasonable society would hear the worries and complaints of everyone within it.
It is kind of noisy because the release recency, which is what your "age" column actually represents, is not important data for the comparison you are trying to make.
Also what message we should get from that table is not really obvious.
> I know artificial analysis quite well as the gold standard in llm evals.
I also know them, but it took me a while to realise you were publishing their data in that table. I don't think it was clear.
> The age is important because new techniques keep being developed and so it is a very rough indicator of the size/cost/efficiency trade-off.
Yes but you are already including the name of the model, your potential public for the table already know about model's release history and therefore each model's age, at least roughly.
There is now an Antigravity CLI which will replace Gemini CLI. Gemini CLI is going to be EOLd by June 18th afaik. Antigravity CLI and GUI share the same agent harness, so it might do the same task.
Well, we were overall better for a couple of centuries after abolishing all-powerful kings + some welfare laws here and there (ymmv, maybe serfdom sounds nice to you). So those changes can work for a while, big emphasis on can and for a while.
Greedy accumulators always end up ruining things for societies when it gets into ridiculous extremes (and there is a part of society that notices and gets fed up).
It is difficult to believe that you can cobra effect yourself into greatness. I'd rather say the most useful perk for companies doing this is the AI-washing adoption metrics they can report, which will hopefully (for them) increase valuations.
I still don't know how to reconcile these reports with what other people say about GenAI-agentic assisted engineering being the only way of working nowadays, especially in startups.
Probably there is no dichotomy going on and it depends on multiple factors, but it seems so weird to see reports that are so different between each other.
It's not required for startups. But if you are building trashy, brittle products and your main metric is speed to market, and have the expectation of high failure chances (e.g. most yc startup batches) - then yes you have to do agentic eng.
If you are making extremely specific, high quality products over a long time window and your founders are deeply experienced in that field of engineering, then no, you don't need agentic engineering and probably want very little llm code in general (outside of some boilerplate, internal toolings, etc).
I think GenAI-agentic assisted engineering is the only way of working nowadays, and it's the only way I personally have worked for months. I still think that an outright majority of presentations on AI tooling I've seen have been in the nonsensical "Look how many tokens I can burn" genre. Had to sit through one guy recently who explained why you need a complex agentic team with 6 different roles in order to ask Claude to investigate a bug, which you most definitely do not.
A guy at my company (very old company, we need to maintain our software for 30+ years) gave a presentation about how they used opus 4.6 to onboard people, like giving new team members access rights etc. Then another guy (even higher in management) proposed using a team of agents for that.. It's getting pretty wild
reply