Sorry Human, You're Wrong (opens in new tab)

Most of the leading models are currently pretty bad at estimating their response accuracy. Difficult questions have confident wrong answers or they bounce around how confident they are (if you express skepticism) to the point where the response is useless. I know it is one of the limitations of how LLMs work but I feel like it is one of the blinders right now with respect to how dangerous the models can be socially and in work settings.

At minimum, the companies should do better to train users on how reliant they should be on their models or how to prompt in a way that provide a rigorous response. Unfortunately doing that reduces the hype around LLMs so they have little incentive to work on that problem.

Found this [0]:

>I've seen discussions about the similarity between Liszt's Consolation 3 (s.172) and Chopin's Nocturne no 2 op 27. I'm not sure if it's just me, but I also notice a similarity between one measure in Liszt's Consolation 2 (s.272) and Chopin's Etude No 1, Op 25. Am I off on this or is this a similar construct?

>Liszt Consolation 2 s.172 approx. measure 52-54 sounds similar to Chopin Etude no 1 op 25 measure 28-29? not the same key but they sound reminiscent of one another, at least to me. Anyone else notice this? Or am I way off?

Looks like Liszt and Chaupin can be quite similar

[0] https://forum.pianoworld.com/ubbthreads.php/topics/3252403/c...

At this point I would have been curious to take a look at the Liszt's composition to understand why the AI thought it was so similar to the one posted. Maybe the AI just found Chopin copied Liszt or the other way 'round... (just joking: I'm not a piano expert in any way)

Good story of why you cannot trust what AI returns. You should check everything.

A confident liar.