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



"Open the Dishwasher Door, HAL." Why Does Imitative AI Need to Be In Everything? [1]

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

Date: 2025-01-07

When I was a teen, my stepmom tried to teach me how to cook. One of first steps was to brown some ground beef. I stood there at the stove and pushed the ground beef around, flipping it and mixing it to get every piece of it brown. I did this for what seemed like hours, but the meat never got brown. It got a nice white color, but never brown.

Did I mention that I am colorblind?

Yeah, my stepmom forgot that, too, because when she came back in to check on me, she asked me why I hadn’t told her the meat was ready for the next step. Turns out, I had browned the meat, I just couldn’t see it because of my particular form of colorblindness.

I bring this up not to allow you, dear reader, to mock my teenage self, standing over a hot stove browning meat that stubbornly refuses to turn any color other than white far past the point where any halfway intelligent person would have asked for help. No, I mention it because Forbes has an article about how appliance manufacturers are adding imitative AI to their products. And I cannot for the life of me figure out why despite being the kind of crappy cook that, in theory, a “smart” stove would help.

The Forbes article is more of a press release, honestly, than a sober look at these potential products. It is deeply, almost sycophantly, on board with the idea of adding imitative AI to your appliances. Wouldn’t it be great if your fridge could look inside itself and see what ingredients were about to go bad and make up a recipe for you! Or order your milk when you are low? Or have your stove look at a picture of your dish and recommend cooking settings and times? Or have your washing machine see your clothes and tell you what should be delicate washed and what should not? Wouldn’t all that be great?!?!?!!? Isn’t imitative AI just dreamy?

Well, no, not actually. These are not great use cases for imitative AI. Imitative AI, remember, has an unsolvable hallucination, to use the industry term, or bullshit, to use the more accurate term, problem. Since these things are predictive calculators and have no model of the world by which to judge real form unreal, they will inevitably produce false information that looks like real information. A fridge that tries to produce recipes will, inevitably, give you something that is at least inedible, possibly dangerous. If you rely on the washing machine rather than the tag on the clothes, you will ruin a shirt eventually.

And that does not even get into the fact that you will likely need to have a wifi enabled appliance for these quires to be run against the imitative AI models.

What puzzles me about these potential products is the why. I understand why Google and OpenAI and Microsoft and to a lesser extent Apple push imitative AI into everything. They have spent a lot of money on these systems, and they are incredibly expensive to run. If they want a return on that investment, they need to hope for widespread adoption at prices that sustain the business. But I don’t see why appliance manufacturers would want to go down this route.

These potential uses all seem like attempts to genuinely make the appliances more useful. Now, the cynic in me also thinks they might be attempts to turn an appliance into a subscription service — have to pay 9.99 a month to get your possibly edible fridge recipes. But my teenage self, for example, would have loved to have had a stove that metaphorically smacked me upside the head when it was obvious the meat should have been browned. But you don’t need imitative AI for things like that — those are relatively simple heuristics. Same with cooking times — they can easily be pre-programmed in for popular dishes. The more fanciful uses do require more than pre-programmed data, but they are also the ones where accuracy is paramount. And imitative AI cannot promise accuracy. So why?

It might just be the idea of using imitative AI as a wedge into the subscription service model. Never underestimate how far companies are willing to punish their users ot make a few extra pennies. But I suspect that this is just another indication that the idea of rational markets belongs on the same shelf as Santa Claus, the Tooth Fairy, and a Blackhawks winning streak.

All of Big Tech is all in on imitative AI, trying to find the next double digit growth market to keep their Wall Street masters happy. Since tech is effectively a monopoly, with only a couple of companies that matter in any meaningful sense, all of the other businesses that interact with tech are going to have imitative AI pushed on them by said companies. And since so many of our industries are monopolies or close to them, it only takes a couple of executives who don’t entirely understand the full implications of imitative AI and suddenly Forbes is writing about how wonderful it will be when your toaster decides it knows how your English muffins should be toasted better than you.

It is not rational in an economic sense — these companies are risking quite a lot of customer goodwill and potentially even legal action for a very limited to uncertain payoff. But is makes perfect sense as a piece of psychology. Big tech pushes something they need to pay off in order to help make it pay off and executives, who very often do not understand the products they make, insist on adding it because the best-case scenarios sound like things they can sell. Add in pressure from Wall Street to not miss out on the Next Big Thing and it is almost inevitable that everything will get imitative AI, whether it is a good idea or not.

Economics should be best understood as a social science. the science of groupthink and bad decisions, perhaps. Things like the push to get imitative AI into everything pretty clearly demonstrate that rationality has very little to do with economics.

Would still like the meat is brown now indicator, though.

[END]
---
[1] Url: https://www.dailykos.com/stories/2025/1/7/2295455/-Open-the-Dishwasher-Door-HAL-Why-Does-Imitative-AI-Need-to-Be-In-Everything?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/