msdz
It appears their end goal is 1:1 drop-in replacements, which other projects usually don’t strive for.
Genuine question, because I don’t know: Wouldn’t it suffice to vendor your current packages and “only” stick to maintenance from there on out, instead of rewriting it all?
This should still achieve the goal of not needing to touch npm, but presumably with less effort/tokens, right? Or is it just that you would expect the result of that process to be less clear and maintainable than a full rewrite?
Wow, that sounds like a lot! Do you use/mean the regular models or flash variants here?
Why not vibecode a Go -> WASM cross-compiler, then?
For chatbot-style usage maybe, can’t really attest to that; for long-running, long-context “agentic” work including planning, within a harness etc., definitely not (yet).
> What I would see even more likely are requirements of no agent commits, use of only human written/vetted code under their name.
I mean in the end that’s what the Kernel project does now, IIRC? Everything has to have a response human’s sign-off.
Yup. Among other reasons, in this case due to the very much nonfictional consequences you’d expect in a “real world” story like this: Had Graw been convicted and stripped of his titles, basically anyone found guilty using evidence from one of his investigations/autopsies, even decades later, could attempt to appeal their case and very likely, this would have completely flooded the courts.
For this reason, this was a huge case here in Munich, and the prosecutors’ office, from what the public could read in newspapers, was very anxious about the outcome all throughout like 2024 to 2026, IIRC.
The tinfoil hat explanation for this is that everyone at Apple is using iCloud for everything, and thus would never encounter a bug with locally-hosted services.
But I should’ve clarified, I meant the scraping aspect of it, since more and more content on the open web will be LLM output.
And with ever-more LLM usage both in labs and the real world, the task of finding new/“creative” out-of-average-distribution material will only continue to increase in difficulty. Meaning if this continues, in the future, all output becomes the equivalent of the color brown?
Exactly. From the page about the requirements for the app under 1) in your link:
> System requirements
> Android 12 or iOS 16
So one of four methods is, by default, already completely dependent on having an account with $american_megacorp.
4) is another app where it doesn’t outright say, but I’m guessing the requirements are also a mainstream OS, mobile at that as well since they seem to use biometrics.
That leaves 2) dedicated hardware and 3) dedicated hardware. Great. (see the “my house burned down” article/thought experiment as to why I’m against that.)
I mean, who cares about the developing minds of the next generation when there’s money to be made involved?
Remember, this is only the company responsible for potentially building and aligning AGI. No need to be setting high expectations toward their sandboxing or hiring capabilities.
That, plus the crowd who would want or like to hack on a device like this, or at least use a non-Apple app or service, is probably far outnumbered by people who don’t care what (and whose) software can run on their screen in the home as long as it looks nice and works well, which is typically the case for Apple.
> I don't see a single objective benefit for choosing hjkl on its own.
“j looks like a down arrow” is easy to remember as a sort of anchor key :-)