wups, /. filtered out my angle-brackets.
should read: ... from "99% sure it's (the right thing)" to "99% sure it's (something not even close to the right thing)" ...

the summary and article here present a delightfully uncluttered surface, but for folks wanting more detail there's a possibly related 2019 paper which shows a variety of image classifiers switching their output from "99% sure it's " to "99% sure it's ", due to literally a single altered pixel in the input. i doubt it's exactly the same thing as whatever this paper turns out to be about but it gives a feel for the problem space.

https://arxiv.org/pdf/1710.088...

> .. But it can't assess authorial intent.

idk if you're using the same popular code-writing tool that i am, but it's way better than human programmers at documenting the whys in addition to the whats and hows. additionally it's a snap to have it curate the architectural and design decisions currently being introduced. the oral record going forward is going to be --much-- more robust than in the past. as far as assessing the authorial intents of existing code, it's pretty good at that too.

i don't love this situation, and i mourn the craft i've spent several decades honing, but it credit where it's due.

i think you're reading a regurgitation viewpoint into what i'm saying. i put 'think' in quotes because i'm still wary of describing them as thinking, but yeah i totally agree they're way beyond regurgitation at this point. i mean, huge portions of human reasoning are language-based. the entire field of formal mathematics is founded on symbol manipulation, for example. so it makes sense that being good at working with language can result in reasoning.

totally,
but where does that reasoning ability come from ?
i'm assuming it comes from the vast body of human discussion about coding.

yeah maybe.

i wonder how much language-independent discussion of software design is out there. can an AI only "think" in a specific language like python or whatever ? i'm assuming the bulk of the code intelligence is coming from scraping SO and all the coding blogs out there, which in my experience are almost always using the context of a specific language.

otoh, maybe that doesn't matter. it can assume python for "thinking" about design and then drop down to the ai-direct language for implementation.

and over time if the bots share their experiences then a corpus of training material will become available. i imagine those discussions, the bots writing their own coding blog posts, will be fiercely guarded by the various corporations producing coding models.

https://commons.wikimedia.org/...

120 years ago..

that sounds similar to the traditional steady-state model. personally i would love for lambdaCDM to lose out to steady-state, but i think there is some inconvenient evidence against it.

previously you mocked me by implying that the bot-forum activity would start once the game started, duh. but now the game is nearly over and there are still only 45 total posts, and you seem to be suggesting that the activity on this thread is somehow relevant. got it. i'm done replying to you.

well, it's now halftime and there are eleven additional posts.

currently there is a grand total of 30 posts.
not sure how this is news-worthy.

C is always so high in these ranking things.
what am i missing ? loads of driver work ? embedded systems ? why would you use C when there's Rust ?

the number of humans is not a function of the "need" for them.

IDK,
i bumped in to the occasional overly-rigid process hiccup at SO but by and large i found it to be a tremendously effective tool, both for asking and answering Qs. imo, the last fifteen years or so of software development owe an awful lot to SO. the rigid rules people are complaining about here (eg closing questions as dupes) definitely contributed to SO's effectiveness.

that said, i'm using past-tense for SO. as much as i owe SO, the many AI helpers are easier and usually better.

it does beg the obvious question of how the AIs are going to get training material for future technologies.

.. looks like a histogram to me. what am i missing ?

https://live.staticflickr.com/...

"1. A vertical bar chart that fails to convey the test score distribution that a histogram would have"