"AI safety" generally has two completely different meanings.

The first meaning, which is what is easy to explain and defensible is that models learn a lot of things that would be dangerous to make easily available to those who search. For example, instructions on how to make bombs. Normally you have to go to school to learn this, as most of the online bomb making "tutorials" are intentionally poisoned. If you try them, you'll end up with something that doesn't work because key step/ingredient is intentionally incorrect.

This is the part that is easily defensible in "AI safety" and the veil behind which people who are pushing for the second meaning hide behind when called on it.

Second one being "political correctness". This is the "trans women are biologically women", "there are infinite genders", "biology is racist" etc. It's basically about pushing the politically correct dogma on The Current Thing.

Both will be sold on the false equivalence of "it's dangerous to talk about child mutilation and castration being bad because it causes trans genocide, as trans people are so insane that they will mass kill themselves if you tell them that castrating yourself and putting on some makeup and a dress doesn't make them women". Which is totally the same thing as making easily accessible and practical bomb making instructions.

And will be forgotten and vehemently denied this was ever the case by the same activists after they move on to the next The Current Thing. While bomb making instructions will remain actually dangerous to society. Which is why AI safety needs to be done on the latter, and not the former. But since activists cannot justify the former without the latter to wider populace, they have to resort to motte and bailey tactics like described above. And that's why "AI safety" became something that is hard to understand. Because just as activsts retreat from their indefensible positions to the highly defensible ones, highly defensible ones become associated with indefensible positions and people begin to question if bomb making recipes being made easily available is actually dangerous, since The Current Thing obviously isn't.

Youtube's bet on the opposite being true for majority seems to have held so far. If you search google for the topic, you'll find quite a few people from large youtube channels discussing their own numbers in light of long term war on adblockers by youtube.

They've all gotten the same impression from their own back end metrics that ad blocker usage has declined massively over last decade.

No, they'll just get a PC.

This has already happened on facebook with images. A lot of image bots were seeded with original facebook memes and such. The end result was that most images were showing some kind of Jeus-like figure with various spiritual backgrounds. Apparently that was the root for most successful viral images on that platform.

It also seems that they were using "any significantly upvoted and commented on image posts" as a metric for their quality control.

So in came the comment bots... and commented on every AI post. And so, image generator AI got increasingly poisoned with AI data that started to grow increasingly incoherent due to lack of sanity checking. It was all the comment and like bots upvoting and commenting on everything making image AI learn that everything works.

And so today, a lot of the AI bot spam... is warhammer 40k chaos spawns with Jesus halos. Horrifically mutated things that have remote resemblance to humans. That is where gradual poisoning of AI data led to over just a few months.

I see no reason why text generators would be different if poisoned like that.

Last elections demonstrated it to be a banana republic, as an endpoint of a decades long process. At this point, US is about the only major liberal democratic republic that allows vote fraud and stuffing at scale through measures like not verifying voters identities, mass mailing of ballots and so on. The main saving grace was the long standing general respect and deep investment that bureaucracy/"deep state" felt for the system and it working well, and so it enforced strict controls and understanding that you don't play with these holes in the rule set or else. That was absent in last elections, as they convinced themselves they were stopping Hitler from coming to power.

And one of the major problems with bureaucratic systems is inertia. Once the "yes there's a way to turn everything to shit, but we generally play it fair because that results in better outcomes for everyone in the long term" tendency was broken in favor of "we must win at any cost", the "any cost" part started to become a norm in everything. This is just another expression of "at any cost" logic that was brewing for a while. Because bureaucratic systems of the state having to interact with many non-government structures rapidly cause the culture in those systems to permeate in everything. It's why more corrupt states tend to have more corrupt companies and vice versa.

While google does screw people, that doesn't mean it's a bolt.

Yes? Did you have a counterpoint that explains how bolts wear out before vehicle is scrapped due to other components failing?

>Plastic clips are not the problem, they're an easy component to source and replace. We've been using them for decades.

Tell me you understanding nothing about engineering and maintenance without telling me you understand nothing about engineering and maintenance.

A lot of things attached by plastic clips used to be attached by bolts because while slightly heavier, bolts last you more than lifetime of the vehicle. Plastic will not, and will get damaged easily.

And difference between "no maintenance needed" and "plastic clip broke off again, needs to be replaced and that needs an hour of work to deconstruct shit to get to it, replace it and another hour to rebuild" is extra spending that is measured in hundreds of USD/EUR every time it happens.

Electric battery fires are a function of damage. It's nearly impossible to assess exact extent of battery damage from normal bangs to bottom of the vehicle. Damaged batteries catch fire spontaneously, randomly and completely unpredictably. That is the problem, there's no such issue for ICE vehicles at any comparable magnitude.

As a result, bottom damage that may have gone unnoticed from the owner because it's sounded like a piece of gravel hitting the bottom of a car actually did enough damage to the battery to cook it a week later. That is what insurance companies have discovered over last couple of years as EVs proliferated.

P.S. The Guardian on progressive subjects has same reliability as Stormfront on holocaust.

Honestly, that's just legacy from PC being an open platform with many operating systems existing for it.

If microsoft could take it away, they would in a flash.

50% in a year is not unusual reduction of battery life for buds, and if you use ANC, there's no way in hell any buds that aren't massive will be able to go 20 hours without charging. Now if you don't use them all that much as you apparently don't, you sure won't care.

For everyone else who actually uses them, yes it matters. To see how much it matters, just look at what is being promoted in marketing.

Apple is one of the worst ones in the bunch for this. They are the ones that pioneered burning down the battery faster in planned obsolescence to sell more new phones.

Oh, that's not a new thing. If you drive older cars (10-20 years old), you know that clicking sound. That fucking clicking sound. That comes from gear wheels in small boxes that open and close vents for car's ventilation system. Specifically it means one of the teeth broke off and it cannot rotate any more, instead just hitting the edges of other teeth making it go "click, click, click, CLICK MOTHERFUCKER" endlessly. Because they changed gears from nylon to plastic molded ones. So teeth break off way sooner.

Some of these small boxes are reachable with just pulling off a panel. Others? You have to deconstruct the fucking dash and then some things under it.

EVs are often charged at home overnight. That means the ever present highly damaging EV spontaneous fire risk is spread to the housing. I'm guessing that is the reason why house insurance is implicated.

While electric cars are by far the worst offenders in repairability, most if not all modern cars have become heaps of easily broken plastic clips with each component having carefully simulated EOL timeline to ensure that cars no longer go for many decades as they used to.

Electric cars just have a problem of lithium ion batteries sitting at the bottom of the vehicle for purposes of lowering center of gravity. Which means that which used to be a few bangs of a hammer for ICE cars is a scrapping order for an EV, as spontaneous fire risks from damaged batteries are intolerable for insurance companies due to extreme amount of damage it causes and time and effort needed to put it out.

The main purpose of insurance is spreading of liability. There will be a certain amount of accidents with certain error margin. All of it needs to be paid by people who use the system that causes the accidents.

It doesn't matter if it's automotive, shipping, housing, etc. The general principle holds. For example in the recent case of Baltimore bridge coming down? Insurance payments will be spread among a massive amount of companies because to spread liability in situation where costs may be astronomical, insurance companies cross-insure each other through insurance clubs. The goal remains the same: to ensure that all damages caused are covered for at reasonable cost to everyone using the system, so that system itself remains viable.