I think there’s generally always been some hackish way around, and the hackish way frequently changes between releases of installation media.

Also, when Windows Update deploys certain things, if you have a system that did this, you may get a full screen thing telling you to set up a microsoft account roughly “for your own good”. Even when you bypass that, the start menu and notification area generally are eager to suggest hooking your system up to a microsoft account.

They decided that since Android can so aggressively push Google accounts, they should get the same scheme going, but Microsoft has this pesky pre-always-online history of being an offline capable OS.

Well auto fellating is the trademark tone of LLM output, which the email probably is.

Sadly I can’t tell if this is a joke or not because I have met so many people who seriously believe things like this work. They are the ones who eventually get the most pissed when LLM messes up on them because they got the LLM to “promise” not to do the specific thing it ends up doing.

They generally evolve their superstitious ritual to something else that will eventually fail, like changing the wording, or making the LLM specifically include a phrase indicating a promise of quality. They also believe when the LLM “apologizes” and think that indicated self reflection and learning. Very few are prepared to accept that the LLM can go off the rails at unpredictable times and unpredictable circumstances, and their utility has to be monitored like a hawk unless the outcome really doesn’t matter.

I say it’s generally a problem of long narratives, but some genres like comedy can get a pass since they don’t have to rely on growth and progression.

To the extent a story needs to develop, running a long time is likely to doom something.

Running a few books or a handful of seasons can work, but if a story has to evolve over decades…

Haven’t gotten around to One Piece (that episode count is… daunting), but I think I really know it’s done as soon as they have a ‘tournament arc’. Give up all pretense and just have them fight for the sake of fighting.

And then there’s bleach, where, oh look, he has a somewhat cool sword, oh it has a cooler form, oh there’s an even cooler form, oh now he has mask powers, but limited, oh wait, we were lying that wasn’t his real cool sword form… Ugh…

Think another example of ludicrous power escalation was when in Loki they just had a drawer of assorted infinity stones. Yes, played for laughs but the problem of escalation suggested is real.

I think the real problem is trying to keep a story going too long, and the need to escalate everything constantly serves to ultimately undermine how that progress feels.

The stories tend to be repetitive, end up where a villain gets a new MacGuffin and the hero has to get some new capability to overcome only for the next villan to have an even bigger MacGuffin, rinse and repeat with each time being portrayed as some impossibly large leap over the last. To keep characters going they time jump, they get cloned, they come back from the dead, they cross over from some alternate universe.

Basically, most genres of fiction have a risk of overstaying their welcome if you try to make it go on a long time.

It’s not so much a problem plaguing fiction in general, but fiction that runs a long time.

If it’s a contained story with defined end that comes relatively soon enough, the stakes can be relatively fixed, arcs can run through to a logical conclusion, etc.

If you have unending, soap-opera like story, then you hit problems. Characters can never actually be fully realized, they have to have their development paused. Any romantic ‘will they/won’t they’ gets ludicrously drawn out. You usually get tougher plot armor because fans are really attached, or a revolving door of characters that you don’t get attached too, or people inevitably managing to be alive after having died. You have power creep where insurmountable challenges get overcome through progress and then something has to reset the new capabilities to table stakes.

Well strictly speaking the full name field is always there, but a lot of people have the full name “”.

But less pedantic, perhaps require was the writing word, but same principle, put whatever you want in dob field, default to 1970 or something.

Yes, the key thing is it might have extracted useful info from otherwise confusing data, it might have mixed up info from the data incorrectly or it might have just made it up.

So it can be useful, if you can then validate the info provided in more traditional means, but it’s dubious as a first pass, and sometimes surprisingly bad when it’s a scenario you thought it would work well at.

Giving us the ow for now, later it’s going to give us the d.

For solutions that back on actually “verifying” the age by requiring credit card or government ID, those suck.

As described, this is an administrator self describing the age, which doesn’t mean much to anyone except kids of people who apply parental controls to systems their kids have access to.

Accounts already require your “full name” but we don’t consider that “full name verification”.

This proposal seems to be in the spirit of least intrusive means to let parents opt into this stuff if they want, with no ties to identity compromising third party/state “verification”.

Question is whether this sort of solution that at least gives parents some chance will satisfy the lawmakers long term. For the wave of laws now, it seems to suffice to self attest age.

Reminds me of a time when my nephew was playing vintage Super Mario Kart and started trash talking me about how long they had been playing and that I didn’t stand a chance.

“I was playing this since before you were born kid”

nVidia had announced that instead of 100 billion for nothing to OpenAI that they were doing 30 billion for stake, and said they were probably not going to keep giving these ‘halo’ AI companies money after this.

I also saw a report that banks were starting to get a bit more stingy with money to the same companies.

I think that while there’s still plenty of money coming in still, it does seem like the ‘take our unlimited cash just because you have AI in your name’ phase is wearing out and they actually have to try to convince people now.

Which is a pretty big problem for them, as despite their brand recognition they aren’t really seen as the ‘leader’ in the AI space on any particular front.

Absolute gigafrood energy.

I find it interesting that, as you point out, his comment makes zero sense but still gets upvotes, and then as he clarifies he’s saying it’s racist against chatbots, then people are unambiguously “oh hell no”

Some people upvote at a baseless accusation of racism without actually actually seeing racism or getting clarification that they may have missed… That’s a bit sad…

Well, it looks like they state three options:

• Passkeys. This won’t work over a medium term, period. It’s tantamount to saying that SSH keys prove someone is human. If there’s enough interest, they’ll just make a software passkey solution that can work. Passkey being “human interactive” is purely a client-side construct.

• Biometric services. Strictly speaking, not an ID but it’s not hard to imagine leveraging capturing biometrics to an ID like scenario.

• Government IDs. Well that’s self explanatory.

They do state distancing themselves from the ID by trusting a third party service, but 3rd party ID service is still a thing.

Of course, this seems to be only after someone accuses you of being a bot and Reddit bothering to pay attention. Which may be almost no one.

Sounds like a statement to say they have a hard-on-bots policy, but in practice it seems they won’t care unless there’s significant outcry against a particular account.

LLMs are not people, they do not possess will, they do not possess sapience. Not wanting to deal with LLM output is in no way like excluding a group of people based on arbitrary characteristics.

Well not really, they added a field so that they could store date of birth in the way they have a field to store “real name”.

So you can be sure my birthday is 4/20/1969 as sure as you can be that my name is Bimbo Baggins.

Note that for the California law at least, this is “good enough” and the OS never actually has to validate anything. In practice a person without admin access could have their birthdate out of control, well, until they run a patched browser that skips asking systemd and just always sends a desired bracket…

It kind of works to keep kids under 13 sending the signal with parental administration, but doesn’t do anything for more resourceful people you tend to find over 13.