selridge

Fuckin WebMD just hunkering down in the corner.

Sure it is. How many trials did we have before ER doctors started using Wikipedia?

You gonna pay for it?

You could absolutely randomize care between a doctor and an AI under an IRB. I’d be stunned if there aren’t a dozen studies doing something like this already.

You have to justify it, but most places have sections in the document where you request review to justify it. It’s not any different from giving one patient heart medicine that you think works and another patient a sugar pill.

Human use is already widespread. You might as well complain in 2015 about the use of Wikipedia among emergency room doctors. That ship has sailed.

The issue is that those hypothetical scenarios do not have to look like how patients actually interact with the tool.

Real life use is full of ill posed questions open ended statements inaccurate assessment of symptoms, and conclusory remarks sprinkled in between. Real use of chat bots for Health by non-clinicians looks very different than scenario based evaluation.

I’ve never heard of in my entire life a doctor failing to recognize a medical emergency. /s

One of the things that people need to come to grips with is that like Wikipedia people will use ChatGPT because it is there. And the alternative is to be rich and have a primary care doctor that you can reach out to at a moments notice. Until that is different people will use these web services. It’s the same thing as Wikipedia or WebMD.

What do you expect that it’s gonna announce itself in a modal dialogue when you run the software?

This isn’t like AI image generation where you’re going to convince yourself that you can tell the difference based on how you think it looks. Do you really think no one in the production chain of any of the software that you use picked up copilot in the last two years?

What signal are you hoping to receive that this is happening?

which textbook would that be?

That's not insider trading lol.

I think that’s a bad faith read on that paper.

What is this nonsense?

You said that none of this was in production and then when people pointed out that it was obviously in production, you shifted the goal post to some other measure that you just imagined in your head.

LMFAO does it hallucinate to the same degree as GPT 3?

Which is what was questioned.

Did google not rely on Gemini to do their ISA changeover?

https://arxiv.org/abs/2510.14928

Was Gemini worse than no tool at all there?

No one is smuggling this in. The debate is over. It's transformative. We're in the midst of transformation.

Fresh take

I sez what I sez

Also worth noting that Apple recently paid a king’s ransom for Samsung RAM

The problem is, we have no real understanding of what people will or will not do with this technology. Will humans only be interested in “real“ activity?

We have no idea, and most people are just guessing in a way that flatters some understanding of art that they have. We also frankly have no idea what the permanent relationship of humans to art is even without AI.

The television is less than 100 years old. There aren’t very many, but there are some people alive today who were alive before the television was created. The computer is about 80 years old. The whole idea of photography and of recorded audio is less uthan 150 years old.

We are still living in the aftershocks of industrial production of art. It is foolish to imagine that in the midst of this chaos, we can point the way forward with ease.

God, thank you.

Finally, someone pointing out all of this is just people announcing what has been in play for half a century.

Yeah, those people love authenticity. They pay a lot for authentic Modiglianis.

Why should anyone care about either of those two people?

Will smith eating spaghetti is art, sorry.

I don’t know that this has to be the way. One thing that is really going to confound this very common idea that taste and quality and personal characteristics will win the day, is that you can use AI to represent all of these to other people.

It’s a huge practical problem to try and figure out authentic nature over the Internet. It’s already clear that people will pay for it, but it’s not at all clear that they will get it. If we imagine that the tools get better and more sophisticated than there is no reason whatsoever to assume that the tools won’t be deployed to give the impression that is needed to make money.

I don’t think any of the above survives if we allow for AI to be used as it is currently being used. It only survives if you pretend that ahead of us is some invisible gate past which this technology will not go.

If we’re restricting ourselves to the United States, then we have to admit that we have been incinerating public education for decades. American schools have had to do more with less and eventually less with less, for as long as most of us have been alive. We have watched in the last 10 years and increasingly in the last five years high school graduates be less and less able to read.

None of this happened because of AI. We could if we want blame smartphones for it, but I think that’s also pretty dubious. We will probably succeed in blaming AI for this. If there is a history, it will get the dates wrong in the 21st-century as to when America lobotomized itself.

We are really not prepared for how few people can competently read and write coming to adulthood right now. It doesn’t matter because we’re gonna speed run the results. Kicking out immigrants en masse means that we can’t even lean on countries that teach their kids how to read and write.

Copilot suggests leftpad

I think we need to invent that distinction, which is notable since the article has MANY opportunities to say it clearly. Instead we are given a picture where the improvement of the agent and the software (here docs are included) is a LOOP, and to make the loop plausible we need to imagine learning in agents that doesn't exist.

That doesn't mean your agent won't improve with a better onboarding regime, but that's a unidirectional process. You can insinuate things into context, but that's not automatically 'learned' and it can be lost at compaction and will be discarded when the session ends. An agent who is onboarded might write better onboarding docs, that's true! But "agents are onboarded mindfully with project docs, then write project docs, which are used to onboard." That's a real lift, but it's best expressed as "we should have been writing good docs and tests all along, but that shit was exhausting; now robots do it."

Don't get me wrong, a fractal onboarding regime is the way. It's just...not a self-improving loop without allowing contextual latch to stand in for learning.

Why would it matter whether or not the robot looks something up if it makes a novel discovery?

Why would it matter that the discovery wasn't just novel but felt like an unconventional one to me, someone who is probably a total outsider to that field?

Both of those feel subjective or at least hard to sustain.

Look. What I'm trying to tell people is that the easy explanations for how these models worked circa GPT-2 is just not cutting it anymore. Neither is setting some subjective and needlessly high bar for...what exactly? What? Do we decide to pay attention to AI after it does all the above? That seems a bit late to the party for cheering on or resisting it.

Some new shit is afoot. Folk need to pay attention, not think they got it figured out already.

"If A and B are separately in the training data, the model can provide a result when A and B occur in the input because the model has made a connection between A and B in the latent space."

This statement (The one I was replying to) is fundamentally unbounded. There's nothing that can't be explained as a combination of "A" and "B" in "training data" because practically speaking we can express anything as such where the combination only needs to be convex along some high-dimensional semantic surface. Add on to that my scare quotes around "training data" because very few people have any practical idea of what is or isn't in there, so we can just make claims strategically. Do we need to explain a success? It was in the training data. A failure, probably not in the training data. Will anyone call us on this transparent farce? Not usually, no.

If a statement can--at will--explain everything and nothing, what's it worth?

I would honestly be surprised if it couldn't do it.

You'd be surprised if an LLM couldn't write *any* program?