> This is incorrect. You don't have to. In fact, you shouldn't.

> There can be situation in which stuff will need to happen regardless, yes, but that does not require lying and probably works even better when one does not lie.

Selling doesn't mean lying. It's usually possible to sell, or persuade, the team to move forward on terrible decisions, without resorting to lying, if you're a good manager.

Having said that, steering a team to execute on good or bad business decisions often do involve withholding information, and some argues -- though I disagree -- that's also lying. If you subscribe to that reasoning, then I suppose you would be right.

Over/undershoots and corrections are of course unavoidable and normal; the absurdity is at the magnitude and rate of change. Furthermore, this is giving it the benefit of the doubt, that measuring AI spend is a good indicator; that's arguably also in dispute. To stretch my car analogy a bit more: it would be like the cruse control system has to hit the target speed, but it only has data from the O2 sensors.

[0] I know that the "classic" cruise control system cannot apply the brakes, but hey no analogy's perfect.

I thought so, too, and perhaps it's just bad luck, but I was at Tokyo Station a few months ago, and I wasn't able to withdraw cash from Mizuho Bank's -- one of the largest retail bank in Japan -- ATM from my US debit card. I ended up walking (getting lost for) ~10 minutes to a Seven Bank ATM, and withdrew cash there without issue. So YMMV.

Therefore, sigh burn those tokens, but make sure your prompts are at least superficially defensible, in the unlikely event that you get audited. Use multiple models for the same prompt / task, for instance. It's well know that LLMs are prone hallucinations, so it's only prudent to double / triple cross-check the results with multiple models.

While it doesn't detract from the article's main point, that Japanese prison conditions are poor, but arson, murder, and jaywalking much? Overstaying your visa is a lot more egregious than the other infractions.

> Damn, I want to move to Japan now.

I know this is sarcasm, but going to Japan as a tourist and _living_ in Japan as a resident -- or the same of any country, for that matter -- are very different experiences. Some, but surprisingly little, of your experience from the former carries over to the latter.

[0] https://www.youtube.com/watch?v=DUHQ_lLD0n0

/s

[0] https://en.wikipedia.org/wiki/NXP_ColdFire

[0] https://www.gnu.org/software/emacs/manual/html_node/emacs/Ki...

----

Like, is nobody in Google multi-lingual? Who the fuck thought this -- not auto-translation, but forced auto-translation -- is a good idea? Surely for an organization that purportedly only hires the cream-of-the-crop, they'll have a larger fraction of employees that speak more than one language? Look, I'm resting-and-vesting like the rest of y'all, but if I were in the team that implemented this, I'd definitely speak up, and let them, up to my skip-level, know that this is terrible. The implication of either possibilities had occurred, yet the feature still shipped, is harrowing.

Even if the developers only speak one language, they must know at least three -- cream-of-the-crop, remember? -- programming languages, right? Imagine if, when you're first hired into Google, you declare your programming language of choice, say Go; then, henceforth whenever you check out the source code, irrespective of its original form, it gets auto-translated into Go, and you can't turn that off? Checking out Pixel first-stage bootloader code, almost certainly written in assembly -- nope! We know better: you're getting that in Go. Fuck, I shouldn't be giving them ideas!

Could they not imagine how horrible this would be, and by analogy when applied to human languages, be also just as horrid?

YouTube's often been cited as a great resource for learning new things. Well, now it's useless for, that's right, learning a second language! I wonder why this Spanish for beginners video's all in English? /s

Speaking about shit features, let's throw "Stable Volume" into the pile. At least this one remembers your preferences...most of the time. When I watch ASMR -- yes I'll admit in public I'm that guy -- videos, and am just about to fall asleep, I just love to be jolted awake by a loud robotic voice's rendition of tapping sounds. Maybe my grumpiness's due to my lack of sleep!

I'd used coffee as an analogy, because it's something that almost everyone drinks, has many varieties, but most generally aren't too picky about it. Alas, I'd underestimated the population overlap between HN readers and picky coffee drinkers. :-P

Yes, but it's like somewhere between mineral water and coffee. That is, most can tell the difference when directly compared, and may even prefer one over the other, but in many cases they are interchangeable.

> Is one generally considered better tasting than the others?

IMHO, no, but you tend to prefer the type you eat most often. Going back to the coffee analogy, most people have a roast / style they prefer, but few would claim that it's better.

> To re-familiarize myself with this bug [...] I downloaded the broken version [...] and tried it out in some virtual machines. Windows 2000 and XP ran it without any trouble on the first try, but Vista and 7 didn’t [...]

Amazing. Emulating an older system in order to debug emulating an even older system. The amount of compute / memory / storage readily available at our fingertips today is astounding. My first computer was a 68k Mac, and back then, I would never imagine such scenarios would be possible!

It's not about wanting non-reproducible builds, but what am I sacrificing to achieve reproducible builds. Debian's reproducible build efforts have been going for ten years, and it's still not yet complete. Arguably Debian could have diverted ten years of engineering resources elsewhere. There's no end to the list of worthwhile projects to tackle, and clearly Debian believes that reproducible builds is high priority, but reasonable people can disagree on that.

This not to say reproducible builds are not worth doing, just that depending on your project / org lifecycle and available resources (plus a lot of subjective judgement), you may want to do something else first.