emodendroket

Sure it is. It's just bound to a vowel.

But that said -- and as others have noted -- the explanation here is misleading, and that's because of the dependence on romaji / transliteration. Japanese conjugation is extremely simple, and the author is missing some essential context that would make it all much clearer. For example:

Doesn't going straight to kana actually kind of obscure the relationship between nomu and nomi that they both begin nom-?

"War is politics by other means" is the one line of Clausewitz everyone can quote.

I don't think there's some magical technical trick that lets you do an end run around politics.

I think there's no question that auditing open-source software is easier, but it can be harmful if auditing actually basically never happens yet people wrongly believe that all the open-source software they're installing must be audited. At that point it's not any better than relying on the fact that technically someone could disassemble binaries to try and inspect them without worrying too much about whether that happened.

I would imagine what it means is that in practice open source just gives a false sense of security that someone must have looked at it.

I would wager a policy of only installing commercial software from well known vendors has a better success rate.

I would say that it's not just an academic argument that's being made about what is technically possible but a stronger claim about what is likely. If the claim is just you technically _could_ do it, sure, that's true by definition.

I know. But the problem is that in reality the only way to get people to audit software reliably is to pay them to do it, so it's not really true as a general principle that open-source software is more thoroughly vetted.

And how often do you do it?

I've seen that people can get Claude and friends to say they're DeepSeek if they ask in Chinese. I think distillation is happening all the time.

A lot of people are slow typists.

No, I've not been "living on" such a principle but it was a big claim for "the bazaar."

I have to say, the principle that open-source software can't do anything nefarious because the source is open just hasn't held up for a lot of reasons -- including that nobody has the time to inspect the code, let alone ensure that it matches the binaries; and also that GitHub has become a distribution hub for software used by lots of people with no ability or interest in auditing the software they use.

Yeah, it'd be easy, but the hard part is getting people to agree what the standard will be.

Plus, frankly, a lot of people (myself included) who want a richer experience are pretty much just happy turning on emacs keybindings in VSCode or IntelliJ or whatever.

But where on the planet are you going to go and escape that?

I'm sure they do genuinely have those problems but it's not like Western research has been free of such scandals.

That's interesting on Deepseek. But I think as long as the models are still making noticeable gains with each iteration it's hard to say "good enough."

Amazon and Uber are other examples of businesses that looked like total basket cases for years. I remember reading, and at the time being persuaded by, articles arguing that Uber was doomed because there are no real economies of scale in livery services, and so the minute they began to achieve a dominant position and hike prices, countless competitors would easily swoop in and undercut them. Didn't turn out that way.

I think there is a real hole in the market for a simple solution that lets you deploy some containers to some instances in a declarative fashion without all of that complexity and does decent LTS versions. I imagine there’s something out there that does this, but k8s has really sucked up all the oxygen.

I mean, it's CDK and whatever equivalents other providers have, isn't it? If you fully embrace all the stuff they give you then it's straightforward to declare everything and it all works together. The downside is the vendor lock-in but unless you actively deploy to multiple environments, which most people don't, you're probably locked in in various ways without knowing about it.

Everyone can decide where the line is for them, but I do feel like given the subject that's a bit of a hyperbolic repsonse.

Yes, I'm sure in Germany nobody ever muffs an interview by offending the interviewer.

I mean, let's be real here. There is such a thing as too honest in an interview.

People are consuming AI like they eat meat: some are embracing it, some are limiting their use of it, and some are avoiding it altogether.

That's an interesting analogy as, despite the real ecological issues with it and principled arguments against meat eating, in general meat consumption has trended upward globally in country after country for decades.

I assume it's because he is seeking to pay rent, food bills, and other expenses through employment.

I picked the guy whose contribution was smaller on purpose to highlight the hollowness of the claim about “controlling the careers” of people who understand what’s going on.

As a brief aside, I am once again extremely disappointed in the myriad of Silicon Valley people who angrily argued that a Democratic led government would ‘pick winners and losers in the AI race’ are now completely silent or defending the actions of this admin. I cannot help but feel that that previous posturing was just a machiavellian play for power, which has just been the worst feeling in the world.

I mean, yeah. But did it take this long for that to be apparent to you?

Yet the guys in the lab coats worked for him.

OK. So? Would you say Harry Truman was a nuclear scientist?

Would you rather they acted like the tobacco companies and downplayed known risks, e.g. all the times LLM output got in the news already for dangerously bad advice, sychophantic encouragement of mental health issues, finding previously unknown security vulnerabilities etc.?

Well, they've done that too, if we're looking for reasons to doubt their sincere concern about it.