kkoncevicius

Similar here. I don't share your numbers, but I had ~10,000 on SO and ~5,000 on Cross-Validated.

I started it as a way to learn things. Mostly picked questions that were on the edge of my understanding, researched the problem for myself, and then tried to provide clear and detailed answers. So I got a lot out of it too.

My view for when SO started to decline is also similar to yours - being flooded with poor quality questions. And on average people asking questions became more arrogant, put little effort in, and demanded the whole place to change so they can paste their broken code and have someone fix it for an upvote.

So I think I naturally and slowly drifted away from using SO. However, publicly stated reasons for SO's downfall often proclaim there was something wrong with people who answered questions. That they were rude or were "gatekeeping". I always feel alienated reading those, since it's very different from how I experienced the site.

What does training data look like for a model like this?

> much of human interactions operates on rules that are not only unspoken but often unacknowledged or even outright denied.

This sounds interesting on it's own. I would be curious to hear more if you are willing to share.

If we could show the current models to someone like Alan Turing, I am sure he would conclude that we have AGI.

Looking at your ladder, made me think where I fit in and I mostly enjoy steps 1 and 2. The implementation and shipping are big chores to me. Once I know, half in my head, that the problem is solvable and I could see a path towards implementation I loose motivation to continue.

So possibly there are other classes of people, not just either enterpreneur or thinkerer. Thou not sure what this combination is, maybe something more like R&D.

And LLMs give a lot of value here. They allow to quickly do steps 2-3 - "hard and tedious" stuff - to confirm that the thing, in fact, became solvable.

The first entries are a bit weird to me. They are mostly about eugenics for people who got their Nobel prices in the early 1900s. The thing is that in those days even the state of California promoted eugenics [1]. And according to Wikipedia, so much so that they might have even played a role in influencing Germany [2].

[1]: https://en.wikipedia.org/wiki/Eugenics_in_California

[2]: https://en.wikipedia.org/wiki/Nazi_eugenics

This was true some time ago but nowadays I don't get this impression. Seems like whatever style I type in, the LLM is already pre-prompted to respond in "its" "own" "style".

Aha, missed it. But unable to delete the thread now.

I had a similar experience. I wanted to test it by asking it to summarise a scientific OMICs-related paper. It gave a warning about me potentially developing a bio-weapon or something like that. And switched back to Opus 4.8.

We can look at the same numbers in different way:

  Error with 91.3% = 8.7%
  Error with 94.5% = 5.5%

  Error reduction = 8.7% - 5.5% = 3.2%

So the improvement is 3.2% / 8.7% = 36.8%

A lot of art from the middle ages is anonymous. Painting itself is an extension of the artist, containing the intension of the person producing it and hence no name is necessary. This is a theoretical state of quality, where activity is not measured in numbers or on a scale but is seen as expression of a particular unique human being. Then comes the renaissance and painters begin to attach names to their works. Here starts a crucial shift - a turn from quality to quantity. Certain artists are better than others and hence quality itself is now measured (quantified) using a name of the person. After that the name becomes so prevalent that some works begin to be valuable only because a certain name was responsible in producing that work. Think - Picasso. Quantity starts to take over. Then comes film and comics and ads where the painter is expected to have no individuality, and he is praised for having a style and technique that is replaceable. Same is true for corporate software development by the way. Here the name (the intermediate state connecting quality and quantity) starts to disappear and is often replaced by a name of a "golem" - a corporation. Quantity dominates - more and faster is better, and the more "nameless" the better. Ten years ago one might think that this is the limit of dehumanisation and it cannot move any further. But now we have AI - where a work of art (or other kind of work) cannot be associated with any quality (cannot be given a name) in principle. And quantity (more, faster, cheaper) dominates. When you think in these terms, the "techno-optimism" is just a place somewhere in this arrow moving from quality to quantity. Or in other words moving from a qualitative anonymity (my work is an extension of my being) to quantitative anonymity (the work is not associated with any being). Hence, it is not a stable position.

Yes! I didn't understand why I always found the Little Prince story (and by extension "alchemist") so repulsive, until I read that book. Little prince is aimed at people who have lost their idealistic youth qualities and seek to get back in touch with that part of themselves. I had the opposite problem - I never fully left that place.

Even if they do not strike gold the second time, there can still be a multitude of reasons:

  1. The innovators will know a lot about the details, limitations and potential improvements concerning the thing they invented.
  2. Having a big name in your research team will attract other people to work with you.
  3. I assume the people who discovered something still have a higher chance to discover something big compared to "average" researchers.
  4. That person will not be hired by your competition.