Ask HN: Will AI = extreme concentration of power?

1 points by atleastoptimal ↗ HN
Let's say that AI will continue to get better without being a major risk due to loss of control. I feel like in this paradigm, it will lead to:

1. Fewer and fewer people can make money via their labor. We will have a small minority of super-winners who magnify their work output 100x with AI, and most people will not be able to contribute, thus will not be able to make money.

2. More and more gains go to those who either own the best AI models (Anthropic, OpenAI, Google), own the infrastructure needed to serve those models (energy infrastructure, datacenters). Everyone else gets access to the benefits of AI only as much as the aforementioned groups cannot better utilize those models.

So far AI models are hard to exploit as a commodity (AI companies benefit more from selling inference than using the inference themselves), but as AI models get better, an AI lab will extract more value from keeping their best models as an internal tool, and sell the outputs of those models. We are seeing the start of that with Anthropic using Claude to make discoveries in bio.

3. AI models become better at capital allocation than humans, so the job of financial planner, hedge fund manager, corporate lawyer, etc. go to swarms of agents, which only the richest will be able to afford. Every loophole, every exploit, every micro-strategy compiled into inscrutable macro-strategy will be utilized by those who can pay for millions of tokens of inference in order to turn their $30 billion into $300 billion. The tactics which humans didn't have the patience, mental energy, or tactical brilliance to employ become just another tool in an infinite toolbox the AI can exploit.

It seems like all these factors spell a perfect storm to an absurdly amplified K-shaped economy, where a very very small minority of entrenched wealth dominates all economic activity, and everyone else becomes a minimally necessary human buffer, kept around for either sentimental or convenience reasons.

8 comments

[ 1.9 ms ] story [ 17.3 ms ] thread
Yes, this is the plan. The top 10% already hold most of the wealth and they are gaining an increasing share, with the top 0.1% gaining the largest share in recent history.

https://en.wikipedia.org/wiki/Wealth_inequality_in_the_Unite...

Any given AI to date has had weaknesses where some humans do better.

If this remains true indefinitely, humans can always fill that niche and be rich.

If it doesn't, you're wildly underestimating how badly it can go wrong for humans, because "things humanity collectively can do which robots controlled purely by AI currently cannot do" currently includes "making compute and robots for the AI".

If AI gets good enough to close this loop, the "minimally necessary human buffer" is "zero"... and that includes the would-be-rich.

That's assuming AI progress isn't significantly curtailed, but just enough such that it only fails to replace the top 0.01% in skill and leverage
I gave two futures, the two uses of "if" splits the future where what you call an assumption is true from the future where it is not.

Also, the "top 0.01% in skill and leverage" are two different groups.

People with a lot of leverage don't only have skill, but also motivation to go that way: the capacity to develop one's skills, without that motivation, directs growth in any of the other complex things people enjoy in life, like being a mathematician or a software developer or a musician.

The end state of capitalism is an ever growing Gini coefficient (disclaimer: I say that without having done any research). Corruption will be rampant and individuals will have the power of entire nation states.

But despite that, the middle class may still prosper. In fact the rich will ensure it, for without a prosperous middle class there will be discontent and uprisings against their oppressors.

So surprisingly, despite a runaway concentration of wealth, capitalism is probably surprisingly stable. As long as technological advances continue, that is. As long as power can be grabbed by other means than zero-sum games such as territorial expansion and resource grabbing. Without that, war and conflict returns en masse.

So with that perspective in mind, I'm happy that AI happened. Without AI, growth was stagnating. With AI we get a few more decades of peace hopefully giving us the time to become a multi-planetary species, save the climate, and solve longevity (longevity probably being important for the human race in order to successfully populate the galaxy, since it otherwise becomes hard to justify and maintain the multi-generational bets of traveling to other solar systems)

So I'm not so concerned about AI alignment as I am with the human alignment of our future oligarchs.

>without a prosperous middle class there will be discontent and uprisings against their oppressors

With AI manipulation, propaganda, enforcement, etc. uprisings won't be an issue