My read of this is "the discussion of taxing wealth makes me anxious. i will do a tap dance, please become mired in watching / discussing my tap dance so that we can put off the inevitable and ultimately necessary a little longer"

To the "conversion rate": maybe, but who cares? The answer here is: apply the tax, see if we still have billionaires afterwards. If we do, then keep doing it.

Maybe it has been a while since you have worked on something difficult, then? Difficult for you, I mean. Maybe the difficult things are now easy.

> If you need unlimited time to write the proof, then you haven’t studied the material yet, and should be expected to perform poorly on the exam.

I mean, maybe. I remember some classes and exams where I had simply never seen the topic before, and the problem relied on having picked up on metapatterns that weren't explicitly taught during the class and applying them in novel ways on the spot. Though my guess was that at least some of those problems were markers for "if you can one-shot this, we want to know".

> If you’re hiring and looking for someone with a minimum of 1 year of React experience then you shouldn’t expect the person to take a full day to do basic tasks with React.

If you're talking about the basics, then yeah I would agree. I guess I don't really know what counts as a synthesis problem in this context, and if we were looking at specific examples it would be more obvious why one might be fair and another might not be. I've heard that line about TAs being touted as being able to solve the tests in just a few minutes, yet somehow unsolvable problems fell through the cracks to the students in the same class more than once in mine.

maybe this is good for weeding out students who would not be a good fit for high stakes speed math.

my semi-superstitious take is that the race to achieve ai is grounded in needing something that knows whats going on and is able to make decisions aligned to generational time horizons. whether that works out or not time will tell, but i get the sense a "good enough" ai is probably our best shot at saving us from ourselves. it's clear we can't do that on our own.

sort of missed the point. market dominance and lock-in means they already are the 800-lb gorilla, and removable batteries sit below where it'd move most people to switch

> The reality is people don't want it, at all.

lmao thats a good blither

https://www.androidauthority.com/removable-battery-poll-resu...

> Also, the lighting connector is better than USB in every way. Mandating an inferior technology is an odd choice.

right, except in the ways that matter and that people care about

you can look at the lightning connector as an example. if you said "if people wanted usb connectivity they wouldn't buy iphones", nobody would take you seriously. and when apple was forced to switch, it absolutely didnt tank their sales because people just loved the lightning connector so much. the bad thing went away and it was great.

If you take instances of softmax that you find in training / inference and there turn out to be a few, and use other things like entmax or sparsemax you see across the board improvements. And like top1 often is just the best answer too, there's a reason why when you're doing tool calls temp=0 is the way to go. Like do you really want creative unicode tokens when writing bash commands. From what I can tell, most of the time softmax is the worst answer that works.

> easier than identifying text distilled from the words of almost everyone alive.

Well, there's more than that going on. AI generated text encodes a high-dimension navigational trajectory that guides the model through its geometry smoothly, like a trail of breadcrumbs. Human speech doesn't do that, it's jagged and jumps around the manifold, and probably doesn't even land on the manifold a lot of the time, and models can recognize the difference pretty quick.