Weird, ICP has always been slop of one form or another, seems odd to get mad about it now …

Nah so their definition is the classical “how confident are you that you got the answer right”. If you read the article they asked a bunch of people and 4 LLMs a bunch of random questions, then asked the respondent whether they/it had confidence their answer was correct, and then checked the answer. The LLMs initially lined up with people (over confident) but then when they iterated, shared results and asked further questions the LLMs confidence increased while people’s tends to decrease to mitigate the over confidence.

But the study still assumes intelligence enough to review past results and adjust accordingly, but disregards the fact that an AI isnt intelligence, it’s a word prediction model based on a data set of written text tending to infinity. It’s not assessing validity of results, it’s predicting what the answer is based on all previous inputs. The whole study is irrelevant.

I guess, but it’s like proving your phones predictive text has confidence in its suggestions regardless of accuracy. Confidence is not an attribute of a math function, they are attributing intelligence to a predictive model.

Why is a researcher with a PhD in social sciences researching the accuracy confidence of predictive text, how has this person gotten to where they are without being able to understand that LLMs don’t think? Surely that came up when he started even considering this brainfart of a research project?

You could make a terrible hash brown …

Man Im not experienced in paying for blow jobs but my gut feeling is that you could get a pretty great one a lot cheaper than that. I dunno, maybe the multi million dollar bureaucracy blow job is really really great.

The Palantirs werent technically evil, they were just stones that let you see the truth, it’s just the strong holders could manipulate what was seen through them by the weak holders to bend the truth seen.

In that respect the company is pretty aptly named, the owners (strong in the rights of the admin) can warp the insights to fit their narrative and feed those to the clients (weak in the rights of the admin). The problem being that the clients are also feeding huge amounts of data (truth) to owners to allow them to do their work. That data is ours.

Arguably, the naming of the company should at least give any government pause for thought about the reliability of the information provided to them and the true cost of feeding information to them. It’s not like the company isn’t broadcasting the issues with their system by using that specific name … surely there’s no in depth discussion of how a system like this could possibly end badly for someone who doesn’t have completed control over it?!

No more recently than that, the reports I was thinking of were more like 2019 or 2023, but as an example here’s one from this year

I mean, I’m pretty sure I’ve heard of recent cases where women have been gang raped just walking down the street as well, but my comment was more referring to ops comment that they should be agitating for better public transport. I agree with the sentiment, but there might be a safety factor pushing the lady to drive. That’s possibly true for walking too. Walking may also not be feasible due to lack of walking infrastructure or mobility issues. My mother in law probably couldn’t walk a kilometre, she’s not obese and is mostly mobile, but she’s just had multiple knee replacements and walking that distance isn’t possible.

Haven’t there been multiple cases of women getting gang raped on indian public transport?

I once watched bake off on ITV and I swear it was 10 minutes of ads to 3 minutes of show, 50min episode took like 2.5 hrs to watch, it was nuts.

If the peer review are unable to differentiate between student output and AI output then they are either incompetent or they are inundated with absolute garbage. The latter also suggests the former is true.

Yeah given the quality of AI outputs they could just read the papers to spot it … you know … do their jobs? I mean there’s a few layers here for thesis review, the supervisor, the professor, the other peer reviewers. They are all supposed to review the paper and at least some of the data that led to its production.

Well that’s a much better question.

Jesus how bad are their student papers that they can’t tell whether an AI wrote one?!

Arguably we should be imposing 25% DST on digital products to counter the 25% tariff on aluminium and steel and then 10% on everything else. The US started it by imposing blanket tariffs in spite of our free trade agreement.