Please consider this my best attempt at forming my thoughts quickly. I think the simplest model is that restricting children's freedom is bad because it restricts the set of all choices they can make, which most likely means they can't make choices that are higher utility (to them). Most of your examples of necessary discipline seem to apply to situations where "no discipline" would restrict their set of choices as adults. Getting hit by a car is the obvious example, but every social interaction and any enforced skill-building time are ultimately in servic... (read more)
I try not to ask people to do things for me that have >50% risk of injuring them, especially when the injury is likely to scar and need stitches, and especially if there is a smaller concurrent risk of permanent injury like levitator ani avulsion* or pelvic organ prolapse. I see rates for the former at 10-30%, depending on the study, and rates for the latter at around 10%. When considering surrogacy, we should include the risk of all possible injury and ask ourselves what ethical weight we assign to that. Cost/benefit goes beyond the price of the contract.
*internal muscle tearing off the bone - surgical interventions are still being tested, so don't count on a 'solution' for all cases
Nice post! Punchy, and I like your writing style. I encounter this problem all the time among tpot/rats. People often like answers more than they like truth, so they ask a question, get an answer, and gleefully glue it into their mental scrapbook without a second thought. I came up with the scrapbook metaphor as I was typing but now I like it. It's like, "look at all these answers I found!"
I think many people, even self-described rationalists, use "evolutionary psychology" as attire. They explain outcomes by appealing to natural selection, but they don't explain how the trait du jour would logically be selected for, nor do they consider whether the same evidence would appear if genes were not the cause. It's certainly more entertaining to sell a story about prehistoric apes' way of life and how it completely explains human society, but I would prefer a story about specific genetic factors or at least a reasonable process of selection which we could theoretically trace from an ancestor to the present.
Just trying to model the possible aftermath of this.
Even if no plagiarism occurred, if training on chats contributed meaningfully to OAI's result, there's still a strong disincentive for any R&D firm against using OAI products. Plus, OAI's behavior (if true) is extremely off-putting. The general culture around knowledge production would shift toward hoarding knowledge, which I think would be a huge loss to progress. I also suspect that many people put all AI companies into the same mental bucket, to where the actions of one heavily influence how they f... (read more)