Do you distinguish skepticism from humility? I'm pretty well aware that I don't actually know anything for certain, and Bayes proponents are always saying there is no 0 or 1. At the same time, for most things, my best decision-making seems to have come from taking my perceptions at face value.
Worthwhile reminder, and education for early-career donors, thanks!
I'd add that if this isn't a one-off for your lifetime (that is, you reasonably expect to have stock in future years, which you might use to fund donations in future years), it's worth setting up a Donor Advised Fund. This lets you donate stock (and cash) without capital gains, but also without having to specify the recipient immediately. You get the tax break in the year you donate, even if you make the donation in a future year. It also simplifies recordkeeping, as all the tax-relevant activity is to your DAF.
"once LLMs write most code, there will be nothing left to do for the people with software development skills".
is a mismatch of quantifiers. If LLMs write most code, there's no need for most of the people with those software development skills that are necessary and which LLMs can do well enough. That doesn't say ANYTHING about the software development skills which LLM's cannot do well enough.
I can't tell if you're just saying "LLMs can't do this part well, yet", or if you're asserting that humans have some ability in assembler that LLMs won't match in the foreseeable future.
Upvote but I'm not sure I understand or agree with the thesis. Programming in C is already pretty niche, and the amount of code that is worth the tradeoff to hyper-optimize (cost to do, to maintain, to re-optimize with new host architectures or microcode optimizations, etc.) is absolultely tiny, and getting smaller all the time.
For most of the tasks where this would be beneficial, the focus has shifted over the last few decades from performance to correctness. The move isn't from C to ASM, but from C to Rust or from validation in C (TLA+ for design, bounded model checker for code).
There still is a place for human optimization based on use cases the compiler-optimizer can't see, but it's small and shrinking.
I haven't seen any papers on this, but I'd expect modern coding agents to write ASM that's more correct AND more performant for optimized subroutines than the vast majority of humans. Really, any optimization for something small enough that you can write benchmarks for and measure improvements, automation is going to win.
I fully support this proposal, but I fear you're ignoring the part that's going to prevent it becoming popular enough for anyone to implement. Decision-makers and populists on the topic of education are focused on the oppression axis, and support of "disadvantaged" groups and individuals, and do not want to accept the model that some kids are inherently variant in ways that can't be applied to all/most.
Personalized/customized programs are generally discouraged for cost and philosophical reasons, and especially so for gifted/advantaged students.
A lot depends on scaling issues - if it's really one in 10k, that's about 7400 kids in the US (there are ~74 million under-18 total). This is feasible to privately fund their education, with some mix of charity, parental payments, etc. Ideally, Robin Hanson's earnings futures would be available - these are great credit risks, if it were legal and acceptable to get them under contract.
But even more depends on the identification problem. Terence Tao wasn't 1/10K, he was 1/10M. Those will almost always take care of themselves - people around them will notice and behave mostly-appropriately. Making it more common to get them into accelerated programs and fund private tutors would be good, but probably isn't the sweet spot for advocacy. The lesser geniuses are less clear, especially early, and especially if there were programs to get them better support and education, such that parents of average+ kids work hard to make them appear to qualify.
I know, and it may be the ONLY thing I know, that I experience myself, and that I do not experience anyone else, except by their effects on me.
Any sophistry that does not acknowledge this is fully disqualified as a search for truth.
I find this easy to believe, but a bit surprising that it's not mentioned or studied or even has crank/subversive pages with POC detections. The printer/scanner steganographic fingerprints became pretty well-known within a few years of becoming common.
I mean, anything that's aggressively online (m365 versions of excel, Windows itself, Google Sheets, etc.) should be assumed to be insecure against state-level threats. But if you've got evidence of specific backdoors or monitoring, that should be shared and common knowledge.
Oh, ok. current levels of "agentic systems" don't have these problems. You can just turn them off if you don't like them. The real issue with alignment comes when they ARE powerful enough to seek independent goals (including existence).
I have to admit that I've never met someone in real life who makes that strong claim. Plenty make the much weaker claim that it's much lower value to create future lives than to reduce current suffering. I personally don't agree with that either - there's no one-size-fits-all valuation of current or future entities.
Fundamentally, rulers and politicians are human, no matter the system or justification. There is no "outside" view, and no passive voice for decisions. Politics is corrupt and hopeless. But no laws can fix it - laws are made and enforced by the same critters as the laws are intended to constrain.
The question is not about what conditions are necessary and sufficient for a representative democracy to work well, it's about what conditions (including what participants) we could likely achieve, and why would that be better than today.