Reminds me of the time iCloud mail was caught silently dropping emails if it contained a banned phrase
(not marking as spam; just completely disappearing them! And giving no indication to the sender, either, letting them believe it went through successfully)
If companies pushing these agents find they're reliably making customers overpay by about ~30%, I wonder why they don't run some sort of subsidy (from their pocket) where like orders placed by your AI are 30% off.
Certainly feels less gameable than the kind of stuff we already see so much of (like referral bonuses, free sign-up balance, "unlimited" promos, etc)
> after the atomic bombing of Hiroshima, some people who would've received a relatively small dose of radiation seemed to have a slightly lower cancer risk than people who wouldn't have received any additional radiation.
What if this is survivorship bias selecting for some common trait that correlates with better radiation tolerance and lower cancer risk (like age).
IIRC, it's based on the distance from the blast radius. The confounding thing about it is the fact that there is apparently a range of distance from the explosion where people have lower incidence of cancer than people who were even further way.
I am not sure what you mean to say with this but reading back through this thread I'm not sure if what is implied here has come through very well. The idea is that well outside the range of where people died from being evaporated, there was a much larger range of people who were exposed to a lot of ionizing radiation from various sources like black rain. But, then, there's a range where people exposed to relatively low doses but higher than background radiation seem to have lower cancer incidence than average. So there is a distance where it is higher than average, then further away from that, it becomes lower than average, before finally going back to average.
This doesn't necessarily imply that radiation hormesis is the only possible explanation, but the easy hand-waves are all basically out. Something weird is happening here and it seems to repeat in different circumstances, too. There is not much to go on other than confounding results and proposed mechanisms as far as I can tell.
The most interesting part about this to me was how they bench/compare it, like in the example with Fable:
"Given the same prompt, Astra for Law returned two closely matching precedents; in the litigation example, Claude Fable 5.1 returned a holding that had been reversed on appeal, while in the transactional example it reported finding no such case."
It made me wonder if a good deal of law is about finding a way to work in statements with clear precedents without your opposition noticing and then later drawing upon them in court (as settled precedents, in your favor) after the opposition (perhaps implicitly) accepted it. That would clarify a lot about why some lawyers need to spend so much time pouring over and memorizing past cases (even ones that are only tangentially related); because anything they miss could be used as a potential trojan horse by the opponent.
If this is true that must mean there are a good deal of cases settled using precedent "gotchas" where both sides knew that without the "load-bearing" precedent the outcome would've definitely been the opposite. (i.e precedents almost always trump even valid arguments)
Confusing enough that I see ads from major retailers mixing it up (or maybe it's on purpose?) listing "Airpods Pro 4" only to click the link and see it's the Airpods 4.
Reminds me of the time iCloud mail was caught silently dropping emails if it contained a banned phrase
(not marking as spam; just completely disappearing them! And giving no indication to the sender, either, letting them believe it went through successfully)
[2/26/2013]: https://news.ycombinator.com/item?id=5287474
reply