pkulak

Yeah, I agree. My policy is much the same as yours: set a passkey only when a site makes logging in with a password such a PITA that the PITA of a passkey becomes the easier option. Google comes to mind. They seem to actively not want anyone to log in. At some point sites will start to 2FA you even with a passkey, but we don't seem to be there yet.

I think the point is that Passkeys are not supposed to be as precious as passwords. You're supposed to have a brand new one for every device/application combination. So they are just login cookies at this point, and haven't even come close to replacing passwords because you need to know your password to get a new passkey every time you log in.

Now, is that the official stance? I don't know; but it's _absolutely_ what every current implementation suggests the companies deploying this stuff want.

You can make passkeys better. Like me, you can install a well-funded password manager (well-funded, because it needs the engineering effort behind it to keep up with the ever-changing passkey apis on every platform in the world; screw up and oos, can't log in today!). Then you have one passkey per remote service, and just have to make sure 1password is _always_ installed and perfectly integrated. Easy!

Yeah, good points. I was a bit snide in my original response, so I appreciate you looking past that. I had even had my coffee by that point, so no excuses!

I still stand by the assertion that coffee is probably, overall, healthy to drink, and at worst neutral. Which is very unique for a substance provably, physically addictive. But I also can't really argue with a decision not to drink it.

So, to summarize:

* People who drink coffee... enjoy drinking coffee, and seek it out, which bothers you.

* Some people (you in particular) have adverse reactions to stopping.

These two consequences _are_ shared by many addictive substances, but also by many other things, like water, or Grape Nuts, or human contact. I really don't see how it makes the thing automatically unhealthy. Like it or not, there's basically no evidence that coffee is unhealthy, and plenty of evidence that it may in fact be healthy.

I wish models called me out more! I can’t count the number of times I’ve had an absolute shit plan, prompted it, and the model built it to a tee. Then I look at the code, and it’s an absolute mess, because it had to make my stupid idea work. Maybe I need a better system prompt.

Same. AI has found so many bugs I would have shipped a year ago.

I notice that the tests are very often for its own benefit. Like, you’ll ask it to stop doing something, literally to just remove code, and it will write a test to verify the behavior is not there.

I can’t imagine a more useless test, but I get that it wants to verify that it actually made the change. I just delete the test when it’s done.

My first thought when reading about this was, why does Pi with 8 custom skills and one extension work so well for me, and apparently so poorly for others that they are compelled to go to these lengths?

I can't tell the difference between Fable and GPT 5.5. I tried Fable while it was in trial $20 mode, used up my whole quota, and it was great, but as soon as I went back to GPT 5.5, everything was the same.

But what I love about Openai is that they still let you hook OTHER harnesses up to a subscription. My Pi setup has been built up for a few months now into exactly what I want and moving over to CC or even Codex is really annoying.

Caveat: I vibe code in tiny little chunks. I see what I want to do, and exactly how I want it done, then prompt that, refine, what was output, then repeat. I bet Fable is better at building a whole app from a 2-sentence prompt; but that's just not important to me at all.

I bet this does a kind of "iframe" thing, where you're really just pulling in full web UIs, and they can be whatever they want. That's the impression I get from the comment about phone clients wrapping the web UI because there's no guarantee about what they will actually be.

My guess is the account was taken over.

Generally not, since the corporate account has all the privacy knobs turned up. I use my personal account on my open source projects, where code leaks aren’t exactly an issue.

Oh great, Idiocracy as a scientific paper.

Careful now, Mr. Goebbels, you're blissfully stomping down a very frought path.

I can’t easily visit my family on the other side of the country for a weekend

And instead of moving or being slightly more careful with your vacation planning, you’re going to do exactly what you want, on the backs of everyone else who has to absorb the negative externalities.

I can't stand military jets over my neighborhood. I get irrationally angry every time I'm walking somewhere with my wife and we both have to completely stop talking for two blocks because some F-14 is blasting around over our heads. They are just unbelievably loud.

This is so exciting.

How exactly? Purely on a tech basis? Because it's not solving a real problem anyone has. Some millionaire who wants to go from NYC to LA a couple hours faster is not a problem worth solving. And certainly not when it's less efficient, and inconveniences everyone in the flight path with a very loud noise.

Those folding clothes clips are the best:

- The robot dragging one corner over itself into a sloppy mess. - Cut. - It’s perfectly folded.

Am I reading that chart wrong? I see Opus ahead across every bitrate.

Yeah, but if you reach for the top shelf every time you need a word, you can't compare things anymore.

And every benchmark is "build GTA-6 from nothing, as a single-page web app".

I think the point is that when you don’t control the ad, it’s a bit tough to normalize its audio. And controlling the ad means bringing ad serving in house, which while possible, is a huge engineering ask.

I guess the solution is to switch to a proper ad insertion company that normalizes to -24 like you’re supposed to, but that’s not cake either. Especially if contracts are signed.

I don’t think YouTube normalizes the audio on their videos. I have no idea why, but that could easily become a quiet video that leads into a -16 LUFS ad that blows your ears out.

I self host everything but email and credentials. Email because I already have a full time job, and credentials because if an emergency is sufficient enough to remove me from my hardware, that’s when I’ll need my passwords the most.

To be fair, I could just sync the db somewhere safe.

That's how it works for every rich/powerful person in every aspect of their lives; maybe to a slightly lesser degree with health.

I'm currently shopping for my son.

Same! I recommend an old (2017-) Nissan Leaf. I got one for very cheap when gas was at a local peak, and now that's it's going down, I'm seeing even better deals. The low range can be a plus when it's a 16-year-old driver whom you have no interest in seeing drive out of the city. haha

it performs on par with a PS5

Wait, really? I looked at the specs and saw like 2/3 the CUs of a PS5.

I bought my last car on Craigslist. It was quite pleasant and I like the car. Not that there isn't garbage/scams there, but there's garbage on marketplace too (more, to my eye). I make a point to always go to Craigslist first. It still exists, and it's kicking!

I pay month to month.

Sure. But OpenAI is the same price. Why would I pay $18/month for z.ai when OpenAI is $20/month?