It's not the tech titans, it's Capitalism itself building the war chest to ensure it's embodiment and transfer into its next host - machines.
We are just it's temporary vehicles.
> “This is because what appears to humanity as the history of capitalism is an invasion from the future by an artificial intelligent space that must assemble itself entirely from its enemy's resources.”
I cannot help but feel that discussing this topic under the blanket term "AI Regulation" is a bit deceptive. I've noticed that whenever this topic comes up, almost every major figure remains rather vague on the details. Who are some influential figures actually advancing clearly defined regulations or key ideas for approaching how we should think about AI regulation?
What we should be doing is surfacing well defined points regarding AI regulation and discussing them, instead of fighting proxy wars for opaque groups with infinite money. It feels like we're at the point where nobody is even pretending like people's opinions on this topic are relevant, it's just a matter of pumping enough money and flooding the zone.
Personally, I still remain very uncertain about the topic; I don't have well-defined or clearly actionable ideas. But I'd love to hear what regulations or mental models other HN readers are using to navigate and think about this topic. Sam Altman and Elon Musk have both mentioned vague ideas of how AI is somehow going to magically result in UBI and a magical communist utopia, but nobody has ever pressed them for details. If they really believe this then they could make some more significant legally binding commitments, right? Notice how nobody ever asks: who is going to own the models, robots, and data centers in this UBI paradise? It feels a lot like Underpants Gnomes: (1) Build AGI, (2) ???, (3) Communist Utopia and UBI.
We don’t know what kind of insecure systems we’re dealing with here, and there’s a pervasive problem of incestuous dependencies in a lot of AI tech stacks, which might lead to some instability or security risks. Adversarial attacks against LLMs are just too easy. It makes sense to let states experiment and find out what works and doesn’t, both as a social experiment and technological one.
They need to be more worried about creating a viable economic model for the present AI craze. Right now there’s no clear path to making any of the present insanity a profitable endeavor. Yes NVIDIA is killing it, but with money pumped in from highly upside down sources.
Things will regulate themselves pretty quickly when the financial music stops.
What is so novel about LLMs (I assume this is the form of AI being discussed) that they require regulation? It’s a dataset, an algorithm and some UI. Almost all the problems brought on by the scale-up are just supply/demand type things. Every problem people point at AI are also problems on some scale with computer software in general, so I’m wary of any regulation (and don’t kid yourself thinking it would be for the people) bleeding over.
Oh they aren't conspiring against democratically made decisions about AI, instead they are "ammassing war chests to fight AI regulation", how submissively worded, but that's expected when they have a grip on all mayor communication channels.
This is why I’ve refused to buy into the argument from these ghouls that AI would make the world a better place, and their occasional lip-service of requesting AI regulation for “human safety”: their own actions paint a dystopian world of mass surveillance, even heavier labor exploitation, the return of company scrip and stores, and the wholesale neglect of human well-being, all while blocking the very regulation they claim to want and/or need to succeed safely.
If these people genuinely believed in the good of AI, they wouldn’t be blocking meaningful regulation of it.
They know that LLMs as a product are racing towards commoditization. Bye bye profit margins. The only way to win is regulation allowing a few approved providers.
Yeah, but we can self-host them. At this point in the span of it, it's more about infrastructure and compute power to meet demand and Google won because it has many business models, massive cashflow, TPUs, and the infrastructure to build expanding on their current, which would take new companies ~25 years to map out compute, data centers and have a viable, tangible infrastructure all while trying to figure out profits.
I'm not sure about how the regulation of things would work, but prompt injections and whatever other attacks we haven't seen yet where agents can be hijacked and made to do things sounds pretty scary.
It's a race towards AGI at this point. Not sure if that can be achieved as language != consciousness IMO
The bottleneck for commoditization is hardware. The manufacture of the hardware required is led by tmsc and samsung being a close second. The tooling required for manufacture is centralized with ASML and several other smaller players like Zeiss and the design of the product centers around nvidia though there are players like AMD who are attempting to catch up.
It is a complex supply chain but each section of the chain is held by only a few companies. Hopefully this is enough competition to accelerate the development of computational technologies that can run and train these LLMs at home. I give it a decade or more.
Another way to win is through exclusive access to high quality training data. Training data quality and quantity represent an upper bound on LLM performance. That's why the frontier model developers are investing some of their "war chests" in purchasing exclusive rights to data locked up behind corporate firewalls, and even hiring human subject matter experts in order to create custom proprietary training data in certain strategic domains.
Every generation has its own copyright war. First file sharing, then blockchains, now LLMs. As long as digital comouters are copy-on-write, this debate shall continue. It will only get solved once we have viable quantum computers.
1) *All major players open source their unobfuscated training data.*
a) The evidence so far shows that every major AI company engaged in intentional and historically unprecedented copyright violation to obtain their training data.
b) LLMs have now poisoined future data for any new players. This is a massive negative externality, and we shouldn't accept this externality as a moat locking out future players from competition.
2) *Levy a 20% royalty on all future genAI revenue to authors and artists who appear in the dataset and exempt genAI from future copywright violations.*
a) The current copyright model is bad for both authors and AI companies. It's hard for authors to collect from violations, and it's expensive and tedious for AI companies to comply with innumerable individual copyrights. Simplify the regime for everyone, and properly reward the people whose work is the foundation of these models.
b) The specifics can be worked out, but, among other things, the royalty should be proprotional to the token count of a work, not just number of works.
Someone once told me about being a new journalist, on the city beat. They said something like: I wasn't surprised to find that bribery was going on; I was just surprised the bribes were so small.
That calls to mind this NYT article about the small time grifts that Eric Adams and associates were up to, including a small acting role in Godfather of Harlem in exchange for canceling a planned bike lane.
As a reminder, Princeton did a study and found that public support for a bill has almost zero impact on whether or not that bill passes [1].
This government is bought and paid for by the ultra-wealthy and large corporations with a system of legalized corruption that began long before the disastrous Citizens United decision [2] but that decision put things into overdrive.
When this bubble bursts as I firmly believe it will, it's going to get much worse because Congress will launch into action... by bailing out those that invested billions into AI without any prospect of ever recouping that money.
I believe that the right regulation makes a difference, but I honestly don't know what that looks like for AI. LLMs are so easy to build/use and that trend is accelerating. The idea of regulating AI is quickly becoming like the idea of regulating hammers. They are ubiquitous general purpose tools and putting legislation specifically about hammers would be deeply problematic for, hopefully, obvious reasons. Honest question here, what is practical AND effective here? Specifically, what problems can clearly be solved and by what kinds of regulations?
The most sane version of regulation IMO is the (already passed) EU AI Act. It's less about control of AI itself, more about controlling inputs/outputs. Tell users when they're interacting with an AI, mark/disclaimer AI-generated content, don't use AI in high-risk scenarios, etc. Along the lines of "we don't regulate hammers, but we regulate you hitting people with a hammer".
Oh wow! I think I found part of the problem. I replied earlier about algorithmic accountability, and the need for Algorithm Impact Assessments, and I got a snarky reply and down voted like I've never seen before. I guess accountability hits a nerve.
38 comments
[ 3.9 ms ] story [ 65.4 ms ] threadIt's not the tech titans, it's Capitalism itself building the war chest to ensure it's embodiment and transfer into its next host - machines.
We are just it's temporary vehicles.
> “This is because what appears to humanity as the history of capitalism is an invasion from the future by an artificial intelligent space that must assemble itself entirely from its enemy's resources.”
I cannot help but feel that discussing this topic under the blanket term "AI Regulation" is a bit deceptive. I've noticed that whenever this topic comes up, almost every major figure remains rather vague on the details. Who are some influential figures actually advancing clearly defined regulations or key ideas for approaching how we should think about AI regulation?
What we should be doing is surfacing well defined points regarding AI regulation and discussing them, instead of fighting proxy wars for opaque groups with infinite money. It feels like we're at the point where nobody is even pretending like people's opinions on this topic are relevant, it's just a matter of pumping enough money and flooding the zone.
Personally, I still remain very uncertain about the topic; I don't have well-defined or clearly actionable ideas. But I'd love to hear what regulations or mental models other HN readers are using to navigate and think about this topic. Sam Altman and Elon Musk have both mentioned vague ideas of how AI is somehow going to magically result in UBI and a magical communist utopia, but nobody has ever pressed them for details. If they really believe this then they could make some more significant legally binding commitments, right? Notice how nobody ever asks: who is going to own the models, robots, and data centers in this UBI paradise? It feels a lot like Underpants Gnomes: (1) Build AGI, (2) ???, (3) Communist Utopia and UBI.
Things will regulate themselves pretty quickly when the financial music stops.
If these people genuinely believed in the good of AI, they wouldn’t be blocking meaningful regulation of it.
https://green.spacedino.net/ai-will-never-create-utopia/
I'm not sure about how the regulation of things would work, but prompt injections and whatever other attacks we haven't seen yet where agents can be hijacked and made to do things sounds pretty scary.
It's a race towards AGI at this point. Not sure if that can be achieved as language != consciousness IMO
It is a complex supply chain but each section of the chain is held by only a few companies. Hopefully this is enough competition to accelerate the development of computational technologies that can run and train these LLMs at home. I give it a decade or more.
All major cloud providers have high profit margins in the range of 30-40%.
1) *All major players open source their unobfuscated training data.*
a) The evidence so far shows that every major AI company engaged in intentional and historically unprecedented copyright violation to obtain their training data.
b) LLMs have now poisoined future data for any new players. This is a massive negative externality, and we shouldn't accept this externality as a moat locking out future players from competition.
2) *Levy a 20% royalty on all future genAI revenue to authors and artists who appear in the dataset and exempt genAI from future copywright violations.*
a) The current copyright model is bad for both authors and AI companies. It's hard for authors to collect from violations, and it's expensive and tedious for AI companies to comply with innumerable individual copyrights. Simplify the regime for everyone, and properly reward the people whose work is the foundation of these models.
b) The specifics can be worked out, but, among other things, the royalty should be proprotional to the token count of a work, not just number of works.
Someone once told me about being a new journalist, on the city beat. They said something like: I wasn't surprised to find that bribery was going on; I was just surprised the bribes were so small.
https://www.nytimes.com/2025/08/22/nyregion/new-york-city-co...
This government is bought and paid for by the ultra-wealthy and large corporations with a system of legalized corruption that began long before the disastrous Citizens United decision [2] but that decision put things into overdrive.
When this bubble bursts as I firmly believe it will, it's going to get much worse because Congress will launch into action... by bailing out those that invested billions into AI without any prospect of ever recouping that money.
[1]: https://act.represent.us/sign/the-problem-tmp
[2]: https://www.brennancenter.org/our-work/research-reports/citi...
https://artificialintelligenceact.eu/
So I'll just say...
Algorithm Impact Assessments
Algorithm Impact Assessments
Algorithm Impact Assessments