AI is already too useful to discard, we could automate 90% bureaucracy and government jobs and no one would feel the difference. But of course fear/angst/rage sells.
People absolutely would feel the difference if AI models started making 90% of the decisions.
You were just denied an application because some LLM decided you fit a certain statistical pattern. Where did the LLM go wrong? What were the steps taken to make that decision? Who can you appeal to? Nobody. AI bureaucracy would just keep chugging along, making decisions based on prejudices and patterns it finds but doesn't feel.
Rainer Mühlhoff makes some compelling arguments in trying to warn people of the dangers of an AI bureaucracy/state.
There are theoretically appeals but for the most part they're illusionary. The original decision is given a lot of deference and the appeal is almost always denied. Plus it's gated beyond a lot of time and money. In adversarial proceedings that's a weapon for one side or the other.
I'm starting to think that when non-experts believe a job will be easy to automate with AI, it usually has hidden elements which they don't understand and which make automating it almost impossible.
Meanwhile, there are all these intellectual jobs which are hard for humans to do, so we assume they're just hard in general. Look closer, though, and many don't involve human social interaction, only require a small amount of good taste, and don't have any physical component.
"...and no one would feel the difference" does speak to the inadequcy of bureaucracy and government execution.
Maybe we wouldn't notice the difference with our interactions with bureaucracy and government, but fairly certain we'd notice the jobs market being flooded with Golgafrincham's cast-offs.
> "...and no one would feel the difference" does speak to the inadequcy of bureaucracy and government execution.
Well, the attempt to claim that speaks to something, but the claim doesn't really, because it's not true.
Let's not even try to get people to think about people who are not them, people who are in need, people with disabled children, whatever. I don't know about the US because it has very few spaces where people actually live together like humans would, but for the majority of the world, at least in major cities: nevermind healthcare or pensions and whatnot, just consider garbage disposal, and maintenance of plumbing and sewer system, the electric grid, and lots of other "small" things like that. The shit we take for granted and corporations moan about because it cuts into their profits and reduces the costs they can externalize slightly. Regulations that say they can't just put saw dust and some heroin in food, that type of thing. That came to be because that's the type of shit they just could not stop doing.
If you could quicksave and experience that just for yourself, without fucking life up really badly for everyone else and for decades, you totally should. I don't think most people who talk this stuff from myopic bubbles would last even a week. Because they don't think so well, we know this from the arguments, and are not likeable, we know this from how they don't consider other people. So to me that's absolute bottom of the food chain energy talking about the Ark B. Pure, pristine, unadulterated projection.
The reason for the existence of bloated bureaucracy is political will not the technical inability to automate processes before the advent of the current wave of AI.
> we could automate 90% bureaucracy and government jobs and no one would feel the difference.
This is really naïve in the long run, since you ultimately cannot solve social problems with technical solutions.
It's always techies that fail to realize the second / third and forth order of events when this level of disruption is measured by the number of mass layoffs you can do in each quarter.
AI is so useful they have to subsidize it with trillions of dollars while having no near term profit. AI is so useful that you have to threaten workers in order for them to use the tools.
I'm sorry but all of this is to be determined but if the present is any indicator LLMs tend to make things distinctly worse, not better. Any areas of "potential" still rely on massive amounts of manual human labor, hardly a trillion dollar industry.
"Useful" in the sense of cheap parlor tricks, useful in the real sense that it enables mass surveillance on the cheap; but actually useless for the material lives of people around the world.
Whereas randomly ranting against "bureaucracy", in light of DOGE and Epstein emails (elite = morons) and what Thiel and the Palantir CEO are on about, is totally normal, isn't projection at all, and talking about "rage" isn't a strawman either. No, that requires a serious rebuttal, being a serious argument and all.
I made one language learning app with a top end LLM backend and at first I thought it was magic. But as I and other people used it more I realized:
- This thing is very consistently lying and misleading people. Do I want to introduce more deception and confusion into the world?
- people don't actually want to use this.
- I don't actually want to use this.
- Something about this feels wrong.
I dropped it. I have another couple of big language learning projects made with 100% human blood sweat and tears, long projects over many years. Zero LLMs or voice models used for anything. Those continue to grow and are loved, and I feel great about them.
Railroads had enormous subsidies, too. This is how infrastructure is built, even in "capitalist" economies because it operates at the level of national security. Even to this day, passenger rail is not profitable, although freight is very profitable. So it wouldn't be surprising if "passenger" AI remains unprofitable, while "freight" AI becomes very profitable.
Pretty bold statement to say it's useless for most people outside of tech. Almost every "normal" person I know including my in-laws are using it regularly. It's becoming the go-to for asking questions rather than Google, Bing, etc.
And the privacy battle was lost 25 years ago. People don't really care if corporations know about their search history (Google), or their private lives (Facebook). You're beating a dead horse there.
I like your analogy on railroads because it leads us to the point that these tech companies should honestly be nationalized at this point as they are a danger to the country writ large.
You were just denied an application because some LLM decided you fit a certain statistical pattern. Where did the LLM go wrong? What were the steps taken to make that decision? Who can you appeal to? Nobody. AI bureaucracy would just keep chugging along, making decisions based on prejudices and patterns it finds but doesn't feel.
Rainer Mühlhoff makes some compelling arguments in trying to warn people of the dangers of an AI bureaucracy/state.