Wait, you lost your 6 digit account number how?

It is well known to developers that writing a new feature is the easy part, debugging it is the hard part. Letting the AI do the “easy part” and having the human do the “hard part” doesn’t really help. Debugging tends to be the largest part of an accurate schedule, and it is the most uncertain part. People frequently make pretty accurate estimates of how long it takes to hit feature complete, but poor estimates of how long it takes to debug those features.

Developers should have an intuitive understanding of this, while product managers, and upper management should have an observational understanding (as in they should have seen it before many many many times and been able to learn it is true even if they have no idea why it is true).

So why is it any surprise to people that having AI write applications and humans debug them doesn’t really save much time?

Maybe, but that isn’t what seems to be going on with AI.

We don’t get the same “news” publication with AI and zero workers. We get a crappy news publication with one worker and AIs cranking out the same volume of articles as 14 workers, but the articles themselves are crap. They manage to get some level of readership and sell roughly the same ad volume per readership level as a “real” news publication. In theory that should show up on GDP because it is a bunch of revenue from ads that didn’t exist before, or maybe that one person would have been 14th of the personal of another publication, so it should look like that one person is 14x more productive.

We get AI in our watch telling us we should see our doctors because we “might” have hypertension. If we do actually see our doctors when we wouldn't have otherwise done so, GDP gets a little boost, one more doctor’s vist! Or it doesn’t get a boost if we move it up 3 months sooner. If we do actually have hypertension and would not otherwise have spotted it before a heart attack and death maybe that has another GDP boost, all the productivity of a person who otherwise would have died.

We also get AI generated hog butchering schemes and mass customized SPAM that slips past traditional SPAM filters. AI seems to be a big friend to scammers, or a big tool of scammers. I don’t know if it is boosting scammers productivity, but assuming it does, how does that get measured in GDP? More fraud moves money from people’s life savings into fraudsters accounts where presumably they spend it in part laundering it into ligitimate seeming income and then buy some luxury goods, or maybe sustenance goods. The people who got defrauded now need to work extra hard to build up a new retirement savings? Do they? Maybe this is an untapped statistic! A major source of GDP growth! We should defraud some people for science! For the dismal science in particular!

I don’t find liquid glass better (although the iPadOS 26 windowing system is much better), but it isn’t exactly like Apple does this every year, or two, or even three or five. They ran with skodomorphic for 7 iOS versions, then with the flat look for 9 versions. I think we are stuck with liquid glass for around a decade now.

There are a lot of details you donâ(TM)t have right here. For example many places require landlords to allow you to have a licensed contractor install a EV charger, and most landlords are fine is you pay someone qualified to improve their property. Second example there are a lot of solutions for charging multiple EVs by hooking them all up overnight and the chargers figure out how to allocate the limited power.

that doesnâ(TM)t mean there are no issues. For example renters are not wild about spending say $350 to improve a landlordâ(TM)s property even if they get use of the improvements for a year or two.

I use to rent a place in CA with a small garage (or really half the garage had been converted into another bedroom). What was left of the garage was the laundry area and tool storage. Car was in the driveway.

The driveway right in front of the garage, which is super common. EV charger ended up in the garage (shared the 30A with the dryer, auto switch that gave the dryer priority when it was on, otherwise the EV got it).

No problem, charge cable went right under the garage door. I guess if someone had wanted to steel $1 worth of electricity per hour they could have done it while I wasn’t parked. Nobody ever bothered to. So I really don’t see “all the junk” in garages blocking EV adoption. It isn’t even a speed bump. Maybe not having a garage at all, but even then if you have a driveway you can make it work.

On street parking is where it starts falling apart. When you can’t be sure you will get to park in front of your home, or if you can’t always do that, if you aren’t “allowed” to run power from your house across the city “right of way” on your own property to your parking space, that could be a problem.

A group of runner women wanting to make sure that they're not fucking a dude that 10 other women are fucking.
Women have rosters of men. But it's not okay for men to do the same thing.

Men seem to care if a women has fucked 10 different men over his life. Women are mostly just making sure the men aren’t fucking ten women right now. Either way I think people owe their potential partners the truth, although for the most part I think expecting people to have had zero or close to zero prior partners isn’t a reasonable expectation while expecting them not to have other current partners is a common and reasonable (although not universal) expectation.

Who in their right mind would post that kind of stuff on a "dating safety app?”

I think the disconnect is you think the women are posting “I’m cheating on my husband and it feels so good!”, they are posting “I think my husband is cheating on me” (and some info about the husband) and getting back “He is, that bastard told me he is single!” (and maybe some info to confirm it, like intimate pictures the wife never saw, or dates he was with the other woman).

In other words it is getting some actual value (assuming the “other women” are telling the truth, which if the various wives have a reasonable standard of proof is going to be not too too hard to establish -- like someone won’t have pics of their husbands in any sort of sexy pose unless the husband at the very least was distributing them with the intent to cheat, maybe they haven’t actually had sex with anyone else yet, but they would have been doing the “check out my horse sized cock, wanna screw” type thing at the very least).

They are, I’m sure TEA has violated laws, but in general a jury is receptive to “no, we totally didn't give this info to anyone we shouldn’t have, some other dude broke in and stole it (online attack == broke in)” so they won’t be heard accountable by a court for violating user privacy, but they WILL be held accountable by potential future customers who for at least some time into the foreseeable future will remember “oh TEA the people that can’t keep anything private? I think I won’t be signing up”, which will likely also result in them not making growth milestones and not getting additional rounds of funding, and just sort of withering and vanishing from the earth.

Which helps future non-users not get their private data hijacked.

It doesn’t help them find out their husbands are cheating, or talk to other women who might be able to get them through tough times by telling them how it sucked for them too but it “got better” or whatever (yeah “whatever”, if I knew what helped maybe I would be better at helping people with emotional issues work through them, but I’m crap at that, so as much as I would like to be the kind of friend that supports my friends in tough times I’m more the kind of friend you go have dinner with and play some boardgames with and don’t really unburden your soul to ‘cause I’m full of useless suggestions and no real helpful whatever it is people offer each other).

Lots of pattern recognition, from quality control to medical diagnosis. Granted that is mostly “machine learning” not large language models, so “so last decade”, but extremely effective.

Basically any problem domain where we can recognize a good solution (and it is unambiguously different from bad solutions), but don’t know any step by step process to get to one have had success stories. Him, no there is more to it then that, it seems like we have spent far far longer not getting to trustable self driving cars then my rule of thumb says. We have made significant progress towards it though, it’s just that “mostly doesn’t drive the car into a lake” isn't a great result...so maybe “& has only a modest failure cost, OR humans don’t do it all that great anyway” is a useful addition.

High by almost a third, Apple’s cheapest laptop is a $650 model of the M1 MacBook Air sold only via Walmart.

CA may have made non-competes illegal, but those are the “you can’t work for our competitors after you leave us” non-competes. It is generally considered valid to prevent someone from working for a competitor while they are actually working for you. Like Apple and Google both say you can’t work for a competitor while employed by them. Both consider each other to be operating in the same markets and thus competitors.

The claim Google/Apple would make isn’t that you can’t work 2 places at once, or even that you can’t in general work for Google and Apple at the same time, but that the employee made an agreement that they would not work for a competitor and lied about doing so. An otherwise legal act which is fraud because you agreed not to do so, claimed not to be doing so, and then did so anyway. So they would certionally have to worry about sabotage, and/or theft of work product. -- Ironically you might be able to wriggle partway out of it by claiming you “weren’t really working, just slacking for both of them at once!”, except neither judge nor jury would really find that acceptable...

an area of the system which you can't look at or audit?

You can probbably look at and audit the PCFax area with some sort of enterprise tool, you just can’t write it. Or maybe they will even allow you to erase it and leave behind a “the PCFax report was deleted, so anything could be wrong with this computer! Odometer tampered with! Tampered!!!”

It is not in HP's interest to instigate more trust into a used computer purchase. This serves to help the used computer market, at the expense of the new computer market.

Maybe.

One reason Apple can charge more money then other companies for new iPhones is people believe they can sell a year or two old iPhone for a reasonable amount of money. HP may believe their products will retain more value over time in the used market then competitors even if the sole reason is because “PCFax says it is good” for HPs and non-HPs don’t get a report because they come to the party late and don’t offer the feature for several more years. Once people get in the habit of buying HP “because you can resell it for like half the price of a new one” the prices of new ones can start to creep up. It would make sense to pay $600 for an HP with an identical spec to a $400 Dell if you believe that in two years the HP will sell for $300 and the Dell will sell for maybe $100, or maybe cost you $10 for an eWaste sticker. Even more sense if HP gives you a 2 year repurchase price because they think the used price will be high enough (i.e. you pay $600 today for the laptop, and it includes a $200 coupon for a 2027 purchase of a new HP if you send the old one back...)

Again the actual HP might not be any better then the Dell (or may be somewhat worse), but it has a PCFax listing that says “great shape! SSD 100% ok! Battery 97% original capacity, great for 2 years of use! Never been dropped! Never had the screen replaced! OEM RAM!” while the Dell’s says “no information, anything could have happened to this pile of junk, buyer beware!”

I’m not even talking pre-orders of video games coming out in a few months.

The entire Chicago board of exchange exists to sell products (originally farm products, you know tons of corn, bales of hay, barrels of hog bellies, whole cows) at future points in time. It revolutionized the farm economy. Farmers can sell this fall’s harvest today in order to afford repairs on tractors they need fixed in order to actually get the harvest into the ground, and keep it watered and weeded and fertilized. Oh, plus afford the fertilizer.

CBOE exchanges billions of dollars of futures a day, granted it isn’t all “farm stuff” anymore. Google buying energy that won’t be produced for ~7 years isn’t really any more astounding then someone buying tons of corn from plants that just went into the ground last week and won’t be harvested until October, and that has been happening since 1973.