Nope, it would be worse. Whether it's made into plastic or fuel won't really affect the supply of plastic or fuel. It won't change the price enough to cause more plastic or fuel to be consumed--whichever one it's recycled into will be produced from the earth a bit less. However if people believe plastic can finally be recycled, they will use more plastic. Whereas if they believe it's burned, I think they will neither be encouraged to use more plastics nor to drive more.

Are you a kind of Utility Monster that feels things to strongly that we should all defer to your moral sensibilities?

I had a (new) table definition where each column referenced a different object field. Those fields didn't exist in the class yet. I asked Claude to create them, using the table as a reference. Claude implemented the first eight and the last two. It looked great at a glance, but there were another 15 fields in the middle that were skipped. I fired Claude.

Were you in such a hurry to respond that you didn't read the following sentence? Businesses care about real costs, not theoretical ones.

I find it unlikely most companies pay for energy used by employees or LLMs. They pay salaries or license them, respectively, so I don't see why the cost of energy to the company is as relevant as the cost of energy usage on the whole (the way it affects everybody).

Keep in mind humans run at ~100 watts, and they need to be powered 24 hours a day, not just when working. There are economic and quality issues, but wasting power isn't one of them.

Caveats: queries do cost much more than 100 watts, but you don't need to run 8 whole hours (a work day) of queries to be equivalent to one 24-hour day (one work day of calories) of human caloric output. Training costs much more than queries. And this is a generous analysis because it doesn't consider weekends. And I'm not saying fire the humans--I'm just talking about energy required to do a job.

Is hallucination equivalent to what we do when we remember something wrong or fail to update based on new evidence? I gather your point to be that hallucination and correct output are the same process. What about for humans? Surely this is also true of us, yes?

Yep. Speed controls and removal of the downvote count may have removed some of the ncentive for content creators to spend more time rehearsing/editing to put out higher quality content.

Regarding "holistically"--on the contrary. When I examine thought process from the "inside", it's a loop of checking associated concepts and applying relevant abstractions. It's also an exploration through solution space. But what I know actually happens during human thinking is much closer to a LLM's operation than my experiential explanation of how thought works. The mechanics are neural networks, while the experience might just be a distraction.

But are LLMs too simple? Yeah. I've always tended to think something was deeply wrong with their architecture, otherwise they wouldn't need so much training. (And I have no sympathy to the argument that humans actually get much more information than LLMs, because that information has almost no entropy.)

Wow, someone is feeling a bit mean today.

Here's a bit more of a systematized formulation of the objection: what kind of behavior would demonstrate that LLMs did have understanding? If there isn't one, I contend that the distinction between understanding and "pattern matching" is one of degree, not kind. I posit further that anybody that thinks understanding is the conscious sensation of something "clicking" lacks the prerequisite philosophical grounding to discuss this productively.

In that case you probably shouldn't have said anything.

We're talking about slashdot markup, not the rendered html.

Maybe they're the same for people that have html rendering turned on for their posts, but they're definitely different for people with other modes enabled.

And for those who'd like to know HOW to see the markup of another post, just reply and click "Quote Parent". You don't actually need to post it.

That, along with the fact that the damage is limited to one building, is why it's a good benchmark. It would be much riskier for these groups to research how to get chatbots to help weaponize pathogens, for example. It's unnecessary, given that the model fails on Molotov cocktails.

May your toilets stay broken!