@GarrettPetersen
I regret posting this because after a cursory search, there's a lot of other people showing the same effect. There was no reason for me to be so skeptical!
That said, I'll stand by the claim that it's probably not produced by the individual rings.
@nearcyan
I think a standardized icon is a good thing. People have picked up by now that sparkle = AI, and so now they if they want AI stuff they know what icon to click.
What's your alternative?
@GarrettPetersen
I'm skeptical. Wouldn't the multiple rings produce multiple projections? Why does it cut off with a sharp edge at the bottom? I think if this is real, the projection isn't coming from the rings.
@goblinodds
This works under the assumption that everyone has some flaw that could be pointed out, creating a mutually assured destruction scenario. But what if Emma were actually perfect?
More practically, I'm thinking about being in group chats with an annoying person who is clearly
@goblinodds
No, I don't think so either! It seems quite mean.
I think what I'm asking more broadly is: is rudeness sometimes warranted? I think the answer may be yes, although I don't think it's justified here.
I'm not trying to find reasons to be mean. I would love to be nice all the
@LinkofSunshine
I read that they often merge into traffic without turn signals because they don't pick up the non-verbal cues from other drivers (honking, flipping them off).
@Noahpinion
Honestly this is a funny rhetorical trick. You can either go "this thing is bad because bad people did it," or conversely, "this thing is good because even bad people did it."
@nathan_cervera
@WallStreetSilv
Credit score is a measure of your ability to pay your debt, not the total debt itself. The US has never defaulted on its debt, which is why investors continue to purchase US treasury bonds and securities.
@RiverTamYDN
Sadly this means they won't have to answer for their statistical modeling crimes. They can just say "well he would have had a good chance if he stayed in, but I guess we'll never know."
@allgarbled
King. I agree, but I feel kind of bad calling them edutainment--Grant is pretty emphatic about how you learn best by doing, literally telling you to stop the video and try to solve things yourself.
@Noahpinion
I very foolishly responded without reading the article.😔 While there are quite a few other big names mentioned (at least from the part of the article I can see), I'd still wager most tech executives and employees are voting Democrat in November.
@JosephPolitano
This happened to me, but it's a school sending me this kid's report card every semester. He isn't my son. Also he's doing very, very badly.
@Genericktag
@FacebookAIslop
Okay, point taken, it is a technicality.
But there is a very good reason why recipes are not copyright protected. There's only so many ways to make a poundcake or croissant--if we granted copyright protection to a recipe, virtually nobody else would be able to publish a recipe
@azimuthbr
I thought this might be it, but I found an original version of the image, tweeted that, then copied someone else's community noted image and tweeted that, and neither were community noted.
@nemontemi
@utkusays
@nearcyan
That's a fair point. The Zoom example is especially egregious, since it doesn't even have text to clarify what "AI Companion" is.
@tracewoodgrains
If it makes you feel any better, I've seen Scott Alexander's posts occasionally get flagged there as well, so you'd be in good company. I don't know what goes into the decision.
It's extremely sad how, when you look for a particular meme on Google Images, the majority of hits are for products: stickers, postcards and other irrelevant garbage.
@vextro9k
@wbhub
@AlecStapp
@teamcoachella
Forecasting is pretty hard! If someone doesn't do a great job by objective measures, but there's literally no one better, they can still be the best. So I'm interested if you can name anyone better. I'm genuinely curious, this isn't just to win an online argument.
@caseytube
@M191308M
@FacebookAIslop
I'm not dying on any hill, I am just stating a very easy-to-verify fact. LLMs aren't perfect, but recipes are an area where they perform extremely well. Feel free to give it a try if you're skeptical.
@skooookum
They should make a podcast that's actually a friendship simulator, where occasionally the speakers pause for you to say something and then they laugh. Like Dora the Explorer for sad grown-ups.
The p-values bow low before our hypotheses, and the confidence intervals shine like the rarest gems. Perhaps Your Highness would care to examine these exquisite scatterplots? After all, the dataset never lies.
@__justplaying
@hyperdiscogirl
To avoid an awkward "no, after you," "No, after YOU," type interaction with the driver, you might pretend you don't wish to cross. The driver then doesn't have to slow down, and you can cross after they've passed.