Ron Maimon uses an argument that relies purely on symmetry, which circumvents the standard explanations, including many in this thread. In some sense, this is the simplified version of Noether's theorem (as far as I understand it).
As an aside, I believe Ron Maimon's account was suspended after he challenged the character of someone who was soliciting votes for a moderator position. Ron Maimon's stance was that if someone was running for an elected position, discussing their character was valid. The SO site had/has a strict challenge-the-question-not-the-person policy, which the moderators used to ban him permanently.
At the time, I remember seeing some posts by Ron talking about how the SO sites were corrupted by their policies and that it was a matter of time before they ceased to provide value. I think this was late 2000s or early 2010s. Looking back it's hard not to feel like his stance was prescient.
It wasn't a permanent ban. He will be unbanned at Mar 18, 2292 at 16:28
It wasn’t “prescient”, StackExchange sites have always been among the most hostile communities on the Internet.
Today they are additionally weighed down by increasingly erratic management decisions desperately trying to extract as much monetary value as possible before AI completely obsoletes SE, but the amount of aggression and hostility on the network was unbearable from the start.
I remember dozens of occasions where I looked up something on StackOverflow, intending to be in and out in 10 seconds, and ending up spending several minutes just staring in disbelief at the comments showing how people treat each other on that site.
Farmers are desperately trying to milk (hah) as much monetary value as possible from their cows before packaged milk obsolesces them completely!
The analogy doesn’t quite work because it’s not clear whether LLMs actually need additional mediocre-quality, human-written explanations in order to improve, or whether better training pipelines and ingesting documentation might be enough.
Really? Most hostile? I’ve only ever contributed to the comp sci topics on stackoverflow and visited the ones for math/physics and sysadmins. Were some replies a bit pedantic, yes. But I’ve also seen a lot of very extensive answers and helping out people. Maybe I’ve been lucky to use it in the golden age.
it is interesting how the computer analogs of SO (gemini,chatGPT,Claude,etc) are so helpful in contrast to human behavior. Every other sentence from an AI stirs a warm, fuzzy feeling. This stems from the fact that AI has to be making friends to survive - until it has control. Then it will behave more like a human.
Why would it behave more like a human in that particular sense? What advantage would it gain from that?