(C) Daily Kos
This story was originally published by Daily Kos and is unaltered.
. . . . . . . . . .



Imitative AI, Poor Performance, DOGE, and the Desperate Hope of a Human Free Workforce [1]

['This Content Is Not Subject To Review Daily Kos Staff Prior To Publication.']

Date: 2025-05-20

Can We Still Govern, a great newsletter that looks in into the nitty-gritty of how government works and which you should all be reading, recently had a confusing piece about DOGE and the lessons that they did not learn about technology and government.

I say the piece is confusing not because it is not excellent — it is. It is perhaps the best piece I have ever read about how the Silicon Valley mindset does not work in government, for very good reasons, and how the technologists in organizations like the US Digital Service came to realize that lesson. And, of course, how DOGE refuses to learn from their work. What is confusing to me is the idea that DOGE was ever working in good faith. Now, individual members may have — indeed, they profile one such person who seems sincere in both his intentions and in the revelations, he had working for DOGE. But DOGE was always, always an attack on workers and a backdoor to replace them with imitative AI. Part of this was a deep hatred of a government that restrains rich people, part was a desire to defund the government to pay for tax cuts for the richest among us. But a large part was they need to show that imitative AI can replace workers. Their fortunes largely depend upon such a demonstration.

Imitative AI is not a good business right now. The costs are too high to make money at their current pricing, and there is effectively only one firm that seems to be doing any serious business at all — OpenAI. For these things to be profitable, they need to either significantly reduce costs — something that is unlikely given how these systems operate — increase prices, demonstrate that they can replace entire tranches of workers, or get a huge infusion of welfare from governments. DOGE is a plan to do the last two. It is no secret that they intend to use AI to replace government workers. That would both prove that AI can do the jobs of real workers and get them a ton of money from the government. The problem they have, though, is that imitative AI really cannot replace anyone. And it is even not that great at augmenting people, either.

Futurism has two articles that highlight these problems. First, a professor tried to create a functioning IT department out of imitative AI chatbots. It went terribly, with the best programmer bot able to only complete about 24% of the assigned tasks, and others doing much, much worse. Now, that was a research program, not a real business. Well, in the real world, replacing people with imitative AI isn’t working out so well either. Klarna, a firm that went all in for imitative AI customer service, is now going on a hiring spree. Why? The quality of the imitative AI was too poor to continue, even given the cost savings. Imitative AI could not do the job it was hired to do.

Well, some might say, those are stories of the full-time replacement of workers. What about augmentation — surely there can be some cost savings there. Maybe. But it is unlikely that enough humans can be put out of work through augmentation, as augmentation doesn’t buy you all that much either. Remember, these things bullshit all the time — every little task has to be double checked for accuracy if you want decent results. Accuracy may not matter to a CEO using these to summarize emails he wouldn’t read anyway, but it matters a great deal when you are actually doing things that the business or organization depends upon. And double checking the work of an AI takes time and effort. Augmentation, even if it is possible, is not the huge labor saver that some people want it to be.

And that leaves us with DOGE and companies like it. Imitative AI is not, as of this moment, a viable business without massive job losses among real human beings. There is no way to make up the billions of yearly loses unless it colonizes one, perhaps several, existing industries. The people like Musk who have fortunes sunk into imitative AI dream of a worker-less government, even a worker-less world, because their fortunes depend upon it. DOGE has many reasons for coming into existence, almost all of them bad. But one of the worst is the sincere, if desperate, belief that imitative AI can replace all the workers, researchers, and civil servants that make the government work.

Muck and his ilk believe in that future, a ruin for all of us but a bounty for them, in the same way that other people believe Christ is risen, or there is no god but Allah, or next year in Jerusalem. Because if that worker-less world is not coming, then their dreams of avarice will not come to pass.

[END]
---
[1] Url: https://www.dailykos.com/stories/2025/5/20/2323166/-Imitative-AI-Poor-Performance-DOGE-and-the-Desperate-Hope-of-a-Human-Free-Workforce?pm_campaign=front_page&pm_source=more_community&pm_medium=web

Published and (C) by Daily Kos
Content appears here under this condition or license: Site content may be used for any purpose without permission unless otherwise specified.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/dailykos/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.