Hacker News Reader

Hacker News

Comment by lowbloodsugar | Hacker News Reader

Full thread

lowbloodsugar

I'll just keep repeating this:

There are three options:

1. AI owned by everyone

2. No AI

3. AI owned by billionaires

If you can make the masses fight for 2 instead of 1, then you guarantee that you don't get 1. If instead, the masses fight for 1, they've got a chance of getting it. You present AI as a false dichotomy: no AI or AI for billionaires. But 2 is a fantasy. There will be AI.

Any of us arguing for (1) get shouted down by the very people who would benefit most from it. The masses do the job of the billionaires.

Most utopian science fiction has AI doing the work and humans leading a life of leisure (e.g. Culture novels). Dystopian futures have AI keeping the rabble under control (Neil Asher's Owner Trilogy, Elysium). Time to choose folks.

View on HN

I wish that those who support #2 looked a lot less like #3.

For that matter, I wish those who were pro-AI were more strictly supportive of #1.

#2 is impossible now that oss models are readily available and nobody would know you are using them.

#2 is not really an option though. It's more like #1 or #3.

I agree with your logic, but you should replace 2 with "AI used by governments only". The haters would have more luck getting rid of nuclear weapons than putting the AI cat back in the bag. Governments will use it for surveillance. Think "sentiment analysis" to make sure you're not a terrorist.

Yeah I'll pick two, thanks.

It's hilarious how the elites don't understand the politics of rejection.

What does #1 actually mean in practical terms? Collective ownership of a giant data center and all the CPUs, GPUs, and DRAM needed to do AI?

This is where the researchers and the essayists should be working. What does it look like when there's a significant tax at every value-add (hey let's call it a value-added tax!) that goes into a big beautiful kitty that funds UBI and general nutrition & housing ?

What?

The opposite I'd say, completely open models/weights with some level of government/civil support for ongoing training (through universities? government labs?), but the infrastructure to run models is private and entirely commoditized/interchangeable with highly competitive pricing from multiple private providers.

4. AI owned by someone else. Unless you’re in America, China, the UK, France or Israel, it’s this or No. 2.

4. regulation... well, that's a no go in the US. So what is the 5th option?

I predict it will get regulated in the US, and that it will lead to regulatory capture. Solving absolutely NONE of the problems people complain about while providing NONE of the benefits AI could bring to society.

#1 seems like the worst possible dystopia. We should shoot for #2 and have #3 as a fallback. The Culture is the worst dystopia I am capable of imagining.

If there are no billionaires, you don't have to worry about option 3.

Reply on news.ycombinator.com