chaorace

I don't necessarily think that this is true. If an AI is designed to optimize for X and self-destruction happens to be the most effective route towards X, why wouldn't it do so?

Practical example: you have a fully AI-driven short-range missile. You give it the goal of "destroy this facility" and provide only extremely limited capabilities: 105% of fuel calculated as required for the trajectory, +/- 3 degrees of self-steering, no external networking. You've basically boxed it into the local maxima of "optimizing for this output will require blowing myself up" -- moreover, there is no realistic outcome where the SRM can intentionally prevent itself from blowing up.

It's a bit of a "beat the genie" problem. You have complete control over the initial parameters and rules of operation, but you're required to act under the assumption that the opposite party is liable to act in bad faith... I foresee a future where "adversarial AI analytics" becomes an extremely active and profitable field.

Genuine philosophical question: if something is created to convincingly resemble humans, is characterizing it as human really an "anthropomorphization"?

Is it the status of the speaker (what you "know" about the subject) or the subject (how the subject outwardly behaves) that matters more?

I know it's a joke, but I can't help but feel endearment for the (obviously not real) guy every time I see this. Who wouldn't want to trade stories with an eccentric foreigner in a silly hat while eating barbeque and watching fireworks?

Technically, that's how all praise is supposed to work. Comparing accomplishments is a short-circuit pathway to tearing others down.

More broadly, this is a consequence of English being a LTR language. The eye scans from left to right when reading these languages, so one naturally wants to place important visual symbols (bullets, controls, etc.) firstmost where they'll offer maximum contextual impact for the proceeding information.

The fixed-width nature of these elements is more of an afterthought that emerges because neat grids make information easier to parse and are generally considered to be visually pleasing. In RTL languages, the movement of the eye is mirrored, so the natural position of these elements also gets mirrored to the right-hand side. Of course, the fixed-width nature does not change because the instinct to craft neat grids exists regardless of directionality.

Recently, however, we have begun to see certain touch-based interfaces adopt right-aligned controls specifically, even in LTR configurations. This evolved to compensate for the fact that most people are right-handed and generally cannot reach across the width of a modern touchscreen one-handed. It's effectively a concession to accessibility and is best implemented as a toggleable feature to avoid disadvantaging left-handed users.

Yes, at least I think so. That's why we so often try to approximate analog with psuedo-continuous data patterns like samples and floats. Even going beyond electronic computing, we invented calculus due to similar limitations with discrete data in mathematics.

Of course, these are all just approximations of analog. Much like emulation, there are limitations, penalties, and inaccuracies that will inevitably kneecap applications when compared to a native implementation running on bare metal (though, it seems that humans lack a proper math coprocessor, so calculus could be considered no more "native" than traditional algebra).

We do sometimes see specialized hardware emerge (DSPs in the case of audio signal processing, FPUs in the case of floats, NPUs/TPUs in the case of neural networks), but these are almost always highly optimized for operations specific to analog-like data patterns (e.g.: fourier transforms) rather than true analog operations. This is probably because scalable/reliable/fast analog memory remains an unsolved problem (semiconductors are simply too useful).

Err... isn't liver doing 90% of the work in that statement? You could just as easily say that "wonderbread, potatoes, and liver" is a miracle diet.

I think the goal is injecting bodies into the surrounding economies, rather being an attempt to substantially reduce the population of Tokyo proper.

-10,000 is trivial compared to 13 million, but even +20 to a nearby farming community is substantial (then multiply that impact by x500).

Preposterous. That line of thinking implies humans are capable of intelligence.

It has been a long time since I've played a round of Civilization[1], but to this day I still default to framing almost every long-term decision in terms of "building tall" vs. "building wide". And then there's EU4, which has drilled into me an obsession with staring a ledgers all day[2].

[1]: Yaddah, yaddah... "Given the opportunity, players will optimize the fun out of a game"

[2]: An unfortunate habit, given the heightened difficulty of pausing IRL time

There's no more productive morning than an angry morning. It's great fuel for tackling the email backlog... albeit sometimes with mixed results.

The thing about git is that it's basically plumbing all the way down. In this case: switch abstracts over checkout, which abstracts over read-tree & checkout-index. The nesting doll just keeps going.

Because switch is built on top of checkout, there's no functional difference between the output of the two subcommands when used for the same purpose (assuming the presence of a skilled operator). It's strictly a matter of ergonomics and abstraction.

At the end of the day, as long as you don't fuck up, just do whatever best keeps you in your flow state.

Worry not, fellow luddites. Basic HTML mode[1] isn't going anywhere.

[1]: https://mail.google.com/mail/u/0/h/1pq68r75kzvdr/?v%3Dlui

Of course, of course. This is why supermarkets are so well known for blaring intense, high-octane dubstep at thumping volumes.

Ok... I'm failing to form the abstract connections necessary to understand the point you're getting at here.

Is the implication that it's inherently unethical to subject others to stimuli under the presumption that the stimulation is uniformly harmless? Surely that's the whole point of studies like this; to rigorously probe such assumptions within the relative safety of a controlled experiment?

I don't really see how that's related, since -- as far as I recall -- that was Microsoft working around outside programs inspecting the kernel, rather than the kernel inspecting an outside program... but please feel free to enlighten me.

I suspect, however, that Windows is a little more sophisticated in their implementation of the concept. "Starts with X" is a much wartier approach than (invented example) a rules-based framework that checks against exe metadata/checksum.

Unforunately, being able to read the document is already sufficient access to make your own copy and alter it however you please.

The best you can do is a digital signature[1], which proves that the document has not been tampered with. Of course, it's up to the recipient to ask for a signed document and to actually check if the signature is still present and accurate upon receipt. Otherwise, it's very trivial for a third-party to remove the signature and add their own edits.

[1]: https://support.microsoft.com/en-us/topic/add-or-remove-a-di...

I see, thanks for the response! I guess you're looking for something more deterministic so you can more effectively pattern match against the error's type?

This kind of bluster is a mainstay of Chinese geopolitics. The law was intentionally written in an open-ended way so that the party always gets the final say on what "completely exhausted" should mean. It serves as a propaganda tool internally and a point of leverage externally -- the party never needs to actually act on the threat for it to serve a valuable purpose.

With that being said, it's not that simple. Bluster isn't just a political tactic -- it also enables the military to slowly chip away at norms and edge closer towards a strategic advantage (e.g.: progressively violating more and more of Taiwan's sovereign airspace). If the party could have things their way, the military would merely continue pushing the envelope until one day a D-Day assault force seemingly randomly washes up on Taiwan's doorstep.

Regarding citation #1: I'm a Rust learner myself, so this is something that's rather salient for me. What do you think about the "boxing errors[1]" approach? This strikes me as one possible way you could bubble up dependency errors without being forced to deal with type conversions.

[1]: https://doc.rust-lang.org/rust-by-example/error/multiple_err...

Worth noting... he says this to contrast his surprise at how well the system works in the paragraph immediately prior:

> I once thought I'd have to weight votes to keep crap off the frontpage, but I haven't had to yet. I wouldn't have predicted the frontpage would hold up so well, and I'm not sure why it has. Perhaps only the more thoughtful users care enough to submit and upvote links, so the marginal cost of one random new user approaches zero. Or perhaps the frontpage protects itself, by advertising what type of submission is expected.

I'm sure dsylexia gets the best of many capable people when banging out URLs.

I suspect that there's also a rather large overlap in the Venn Diagram of "people who use Github regularly" and "people who are sufficiently mentally toasted as to regularly make implausible errors".

You've inadvertently described the delineation between a DE and WM. WMs manage windows and that's it -- they're lacking by design.

A DE (e.g. Gnome, KDE) is a multi-part package that bundles up a given WM with the software suite necessary for running a conventional desktop. There exist plenty of DEs that bundle up tiling WMs for convenient use (for example: Regolith[1])

[1]: https://regolith-desktop.com/

There's also an English implementation of this same layout concept on Android and iOS called MessageEase[1]. I've been using it for almost 10 years now and have been quite happy with the support over time.

You can consistently type without looking at the keyboard once you learn the layout, which I've found to be great for note-taking

[1]: https://www.exideas.com/ME/index.php

I wonder if FDR would still be in office if he were still alive? It's an interesting thought experiment to consider how long he'd have lasted given his prior record.

Ah, the one the epoch is based on?

Makes sense. So much of the work around CERN already happens independent of whether the collider is running or not, so skipping a rough season seems like the sensible thing to do (even from a purely inward facing cost/benefit standpoint).

It's a shame that the timing just had to coincide with the start of LHC Run 3, though. They only managed to generate around 5-6 months of data after 3 years of upgrades. I imagine that should still be enough to keep the scientists busy this Winter, though.

QMK firmware has a handy feature for this called Tapping Force Hold[1]. If you hold the key normally, it modifies the layout, but if you tap it once (or more) and then hold, you'll get the normal repeating behavior.

Example A:

Hold Backspace => Activates layer overlay

Example B:

Tap Backspace => Input backspace

Hold Backspace => Input held backspace => OS handles repeating as configured

[1]: https://docs.qmk.fm/#/tap_hold?id=tapping-force-hold

Color me skeptical. Demand for a pint isn't about to disappear just because energy costs are rising. Surely the more likely outcome is that pubs will pass along rising costs to consumers, who will absorb increased prices because: *A)* pubs are a centuries-old English social institution with no readily available alternative and *B)* alchohol is habit-forming.