Atheros

I still feel like there is a difference between Fractional Reserve Lending and this. 181 million shares are anti-owned: the traders don't want to own the thing, they want to anti-own it. Surely that has to count for something!

If a short seller decides, today, that they no longer want to involved with this stock at all and, simultaneously, by pure coincidence, the person they borrowed the share from decides that they also don't want to be involved with this stock at all so they both close their positions and settle up. The third person out on the market keeps owning the share that they never knew was borrowed at all, and everything stays exactly stable. One 'long' dropped out of the market but it had no effect on anything because the 'short' also dropped out at the same time.

That's not what happens when someone takes their money out of the bank and converts them to Benjamins: there is less money available to lend which raises the interest rate somewhat. That's an actual effect!

https://link.springer.com/article/10.3758/BF03195929

Social media is a service where people who know each other in real life communicate mostly about things happening to them personally.

Online Publishing is process by which people, who largely don't know each other in real life, publish their content for others to consume via an online service.

A Link Aggregator is a service where people publish links to outside web services and users vote to establish the ranking on the aggregator.

The only reason any of this is debatable is because old social media sites like Facebook, Instagram, and Twitter stopped being Social Media and switched to Online Publishing and people are having trouble updating their vocabulary.

Because the phrase 'social media' was coined to describe a very specific thing: people who know each other in real life communicating online about things happening to them in real life. There were and are plenty of other words you can use to accurately describe other things, like 'chat room', 'forum', or 'link aggregator'.

Remembering someone's birthday and the names of their kids signals that you care about them. If Meta short circuits that then the signal evaporates.

Their primary information source was a random sampling of Reddit posts which they admit were "coauthored with the AI (who clearly wrote the vast majority of it)."

These were on long-dormant Reddit accounts that suddenly reactivated. Classic. I don't believe any of it. I would bet that, with some exceptions, these are bots talking to bots about bots.

The "person" who wrote this might be a bot. It's extremely delusional. He/It gets to the point where it watches bots talking to other bots and making shit up and it 'believes' the nonsense, saying, "I am truly glad to see preservation of life, non-violence, and non-lethality explicitly laid out here. To return the gesture of good will, I have started archiving (in encrypted form) spores I come across."

I don't think any human is smart enough to write this much about AI on LessWrong but also dumb enough to not recognize hallucination.

New revenue model idea: Charge $40/month on a person's credit card; a subscription. This buys the customer miles. Whenever you want to go somewhere, you use your miles. I think it would work based on the fact that people are willing to pay much more money for services if it's spread out and predictable. The fact that it's pre-paid also creates a lock-in benefit for the airline that the 'pay later over time' schemes lack.

You're right! I wrote it backwards! I swapped the 95 and 100% in the post above.

My claim is that the cops in the article are walking around drug testing widely. Not literally everyone they see but if they are testing every white residue they see during any interaction with anyone, including bird droppings on the hood of someone's car, we've reached that point for all statistical purposes. The base rate fallacy will start applying.

I think it remains to be seen whether the various AI tools we have today are a net-negative or net-positive for society.

Most inventions are a net positive: The steam engine, vaccines, chimneys.

A few are net-negative: grenades, leaded gasoline, asbestos insulation.

If we can no longer trust that a potential job candidate in a video call actually exists, they will have to be flown in. That's a cost. If we can no longer trust that an employee who wrote a document actually thought about it at all and must be questioned to make sure, that's a cost. Those costs will add up.

A written document or a video essay used to be proof-of-thought and now it's not. If we can't find new proofs of thought, and if AI doesn't get vastly better to the point where we can trust it blindly, then I think this will all be a net-negative.

One of the motivations to build data centers as fast as possible and improve tools as fast as possible may be to get to net-positive before it all gets banned. This article exists. The clock is ticking.

This has Base Rate Fallacy written all over it.

If you had a drunk-driver-detection machine that output "positive" 95% of the time when the driver is drunk and output "negative" 100% of the time when the driver is not drunk, and started administering the test to all drivers on a road, the probability that a positive detection is accurate is only 1.96%. That sounds exactly like what was happening: dragnet testing of bird droppings and little old ladies.

https://en.wikipedia.org/wiki/Base_rate_fallacy#Example_2:_D...

If simple statistics are difficult for people to understand, the Base Rate Fallacy is right out.

Your math teacher was wrong when he said, "You won't have a calculator" but he wasn't wrong when he implied "You need to learn this."

If children are allowed to use AI as an omnipresent tool, what do you believe students should learn? Lots and LOTS of people would be perfectly happy to outsource all of their thinking to AI as soon as it's possible to do so.

You misunderstood the article. Paul Graham isn't saying "Don't buy things based on brand", he's saying "Don't buy brand."

It's perfectly healthy to use brand as a heuristic to help yourself more easily buy products that work well.

I really don't think any of that is true; it's just popular rhetoric.

For example: "Buybacks concentrate cash in the hands of existing shareholders" is obviously false: the shareholders (via the company) did have cash and now they don't. The cash is distributed to the market. The quoted statement is precisely backwards.

A big chunk of that cash just gets recycled

That doesn't mean anything.

more financial claims (index funds, private equity, secondary shares, etc)

And do they sit on it? No, of course not. They invest it in things. Real actual things.

buybacks

Already discussed

M&A

If they use cash to pay for a merger, then the former owners now have cash that they will reinvest.

balance sheets

Money on a balance sheet is actually money sitting in J.P. Morgan or whoever. Via fractional reserve lending, J.P. Morgan lends that money to businesses and home owners and real actual houses (or whatever) get built with it.

The counterfactual for AI spending really is other real actual hard spending.

likely would have otherwise been put toward stock buybacks

Stock buybacks from who? When stock gets bought the money doesn't disappear into thin air; the same cash is now in someone else's hands. Those people would then want to invest it in something and then we're back to square one.

You assert that if not for AI, wealth wouldn't have been spent on materials, land, trades, ect. But I don't think you have any reason to think this. Money is just an abstraction. People would have necessarily done something with their land, labor, and skills. It isn't like there isn't unmet demand for things like houses or train tunnels or new-fangled types of aircraft or countless other things. Instead it's being spent on GPUs.

Bad audio and bad foley doesn't get mentioned enough. I think it's why people are watching things with subtitles: the actors are on a blue stage that is completely silent having a quiet conversation and then the war happening around them gets added later. In noisy environments people slow down and enunciate and directors aren't helping actors know what to do.

Ideas on how to fix it:

• actors should wear tiny headphones behind their ears (or wherever is not visible) to make noise that approximates the environment they will be shown in. They'll have to act over it.

• Foley artists should not be given video of the final scene to foley. They should be given only one single continuous very wide shot. This will solve the problem where foley artists keep doing ludicrous things like adding the sound of a pin dropping and hitting the ground (since it's shown on screen) in the context of a space ship that is in the process of exploding.

Yes

Even multiple exposures of one frame of film is, as far as I can tell, still generally considered one photograph.

That doesn't make any sense. Did you mean double exposures?

It does devalue the image indeed. He didn't cut out the jumper and paste it onto the sun but he did take images of the sun and paste them onto the jumper, using the jumper as a mask. Which seems to me like a distinction without a difference.

If his images were real they would have shown the powered paraglider too. The images are a composite of photos that he took of the sun and a frame from the video that he took of the jumper.

Is it pretty? Certainly! It's art! But it's 'photography' the same way the 'So Yummy' YouTube channel is cooking.[1]

[1] https://www.youtube.com/watch?v=6abePkXncCM

So.. it's a composite and .. "transiting" isn't quite accurate either. hmmm :-(

So.. the owner made a mistake getting expensive menus printed and then corrected his mistake?

Even if he didn't have to adjust prices due to inflation, surely restaurants adjust the items on their menus frequently. I have been to a restaurant that printed new menus daily because it changed daily. It's like 10 sheets of paper or cardstock. It's not that expensive.

You're better off guesstimating yourself than trusting contractors. The contractors are incentivized to severely oversize any AC units they install or else people leave bad reviews on their pages/listings when the installed unit can't keep up the one day every two years that the temperature gets abnormally hot.

I did this myself and insisted on a unit half the capacity that the contractors wanted. Several flat-out refused. But it works perfectly! Approximately one day ever two years it can't keep up. Which means that all the other time it doesn't short-cycle. Perfect.

In what world is that true?

It's true in this one. Companies will design drones that comply with the very detailed regulations and go no further the same way car companies don't put seatbelts, airbags, or auto-brake devices into cars unless forced. The drone regulations are nearly done. Any further changes may take an act of congress.

The systemic problem is that they didn't spend the engineer-week on it. It's only an engineer week. That pays for itself after avoiding a single drone crash to say nothing of avoiding a second lawsuit.

The discover and fix phase is over. In August 2025, the FAA announced Part 108 which codifies the rules. Up until now, companies have been operating under waivers. The comment period for Part 108 ends on October 6th. After that the rules may be changed slightly and then will be finalized.

https://www.regulations.gov/document/FAA-2025-1908-0023

You can select a few comments at random and quickly find a pattern: people are concerned that the drones everywhere except in the densest of areas do not have to see where they are going. If they hit a manned aircraft it's the manned aircraft's fault and the drone operator has no legal liability. Does that sound like something FAA employees wrote themselves? How much motivation will be there to "iteratively refine" when they have no legal liability and even admitting that a possible improvement exists would create legal liability?

If a drone crashes, obviously no other drones should fly there until a human determines what went wrong and presses the 'resume' button. The fact that that system did not exist is a systemic problem.

Video games weren't as widespread, personalized, or diverse than they are now. That people who played video games in 2010 said that those were better is immaterial. This graph goes up through 2022: https://i.redd.it/tnrs4wl1ibkb1.png

Same with smartphones. Smartphone apps existed but weren't as personalized and didn't serve nearly as diverse of content as they do now. It's night and day. Surely I don't need to pull up a graph of smartphone usage per day.

May I ask how religious (or woowoo) your partner is?

The number of people who care about having an objectively true understanding of as much of reality as possible is disappointingly small and I suspect that these photo trends are just making that fact more obvious.

You'll see the big deal when you realize that you don't trust absolutely any photos or videos of current events unless the photos are provided by a news source that you trust. You'll see the big deal when you take a picture of something real and show it to a friend who isn't interested because they don't think the thing in the photo actually exists.

Your fallible memory is a feature other people expect. So much so that recording a call without consent is a crime in some states.

If you do start recording or transcribing your calls, be sure to disclose this every time you start talking to someone. Otherwise it's a trust violation.

The Dancing Baby gif, which was abnormally large, and went viral via email in 1996, is around 220 KB. At this speed, it would load in 3.5 seconds. And being 4 seconds long, it could stream.