It seems "Mythos is really good at finding vulnerabilities" has been what people took away from the Project Glassing announcement, which makes sense. Unfortunately for Anthropic, most seem to have forgotten the best argument Anthropic had for holding Mythos back from the general public, "it's crazy good at crafting exploits". Then, without that context, the tinfoil hats came out.
> Introduce and enforce structs for passing context and input shapes around. So as to stop fighting with NULLs, lack of keys in maps and other maddening cases that inflate your coding lines for no other reason than programming languages not having higher-order constructs on well-researched and mostly resolved computer science problems
Once upon a time, I thought I wanted to learn cursive handwriting. Except this version of me was already in his 20s, would be out of school in a matter of months, and quickly realized the skill would be of such marginal utility in the future that it wasn't worth the hours spent tracing out giant letters like a kindergartener every day.
One could learn this skill in their 20s or beyond, but there's an opportunity cost – why not something else that would actually improve work performance, or that you enjoy doing?
I still wish I'd been taught in elementary school, though, because it would've been really useful as a student. Some of our teachers discreetly handed out practice booklets to students who'd "expressed interest" (their parents taught them the basics and teacher noticed); most of us were not so lucky.
Does "applying knowledge" necessitate human-like intentionality and theory of mind? If you insist it does, and this is a category error, then we need a new category.
By analogy, consider that many have referred to classical, deterministic computing as some kind of "thinking" for the last half century+. Does this stop being kosher when the computer has an uncanny propensity for human language? Perhaps, but the computer is still clearly chewing through problems that would have required a lot of human thinking (e.g., arithmetic) in ages past.
I haven't seen any genuine proposals for words to replace the human mind analogues, let alone proposals that the anglosphere would plausibly adopt en masse.
Claude in general probably increases observed bugs in rsync, because it can churn out vulnerability reports that necessitate tons of changes to software that people are accustomed to working flawlessly in non-pathological use cases.
I don't have empirical evidence for this claim, but best I can tell, security patches are the principal source of observed bugs in software of a certain vintage, because they cause churn. (Just think of Windows updates that break drivers.)
The baffling part is why it takes hours for the npm security team to unpublish packages that contain malware, as attested by multiple independent sources? That should be able to happen in minutes.
Postinstall scripts have remained an effective attack vector for quite a while – which, ironically, has meant the worm's authors had little incentive to try something else, so it was easier to inoculate yourself. Alas, you're right, it should be pretty simple to bypass this kind of protection, if they haven't already (and seems like they have).
Can't speak for intelligent autocomplete writ large, but I treat it as an ergonomic feature, and Cursor's implementation is pretty good (though I'm not sure it's improved all that much in the past year).
It constantly takes whatever is currently visible in your editor to feed its context. If you get a nonsense/hallucinated suggestion, you can accept it, get it to read the error message from LSP diagnostics, undo, and then it'll correct itself next time. Or if you need to make changes in 5 places, and the next 4 changes are easy to guess after seeing the first one, it'll guess the next 4 for you.
I still use standard IDE features extensively. The intelligent autocomplete is just another tool to reduce typing when the next change is easy to guess.
Oh, and I turn it off when I'm writing prose or need to actually think deeply. Then it really does hurt more then help.
(Worth noting: I currently work primarily in Go, which is a language that's ridiculously verbose and has lots of repetitive patterns. YMMV for more expressive languages.)
> To avoid this, you will probably need to intentionally write in a very different style than you usually do (or to have AIs rewrite all your prose for you, but, ugh, that’s not a world I look forward to living in).
I agree. The amount of vague and cliche'd AI writing I read on the daily is already exhausting enough.
It would be interesting if you could train a model to sprinkle random red herrings throughout your text in a minimally disruptive way. But I fear you might have to stretch the definition of "minimally disruptive" to make it robust against detection.
The (shift+)cmd+` order also resets to match the window z-order whenever you switch apps. So if the order is windows A, B, C, then you select window B, cmd+tab away, then cmd+tab back, the order will now be B, A, C.
I've developed an intuitive understanding of this, but I had to experiment just now to describe the behavior precisely. And my intuition is still wrong sometimes (like if the app has windows on multiple monitors, it's hard to predict the z-order).
> if I Slack open in Firefox in workspace 1 and Outlook open in Firefox in workspace 2, there is no way to switch between Slack and Outlook
My local maximum is to never use workspaces – just cmd+tab, cmd+`, and sometimes cmd+h to reduce screen clutter.
What can Bank X do to stop phone malware from scraping the user's session token from the Bank X app or website?
Yes, banks should (and sometimes do) double- and triple-check with you before allowing large transfers/withdrawals, but scammers know how to coach their victims past this. Speaking from experience.
(I also don't fully agree this is Google's responsibility, and I am not happy about this development. But there are legitimate points in favor of outsourcing the question of "will this software do nefarious things" to some kind of trusted signing authority.)
Wood is edible when processed correctly, but it's not legally considered "food" because there are a bunch of nontrivial steps to get it into that state. Likewise, any reasonable interpretation of "general purpose computer" in this context by a judge would not include your microwave oven just because someone with skill and finesse could transform it into a cursed Doom arcade machine.
Laws are interpreted by people trained to fill in the blanks[1] with a best guess of the legislative body's intent. And the intent here seems pretty clear: to regulate computing devices that let end users easily install software from a centralized catalog.
[1] which we all do subconsciously in day-to-day speech, because all language is ultimately subjective
The irony is that web searches for an explanation of something often lead to a discussion thread where the poster is downvoted and berated for daring to ask people instead of Google. And then there's one commenter who actually actually explains the thing you were wondering about.
It's kind of nice, though, because you can click anywhere on a window to focus it. If you want to interact with a background window without focusing it, hold Cmd and click.
This is already the pre-26 bounding box, isn't it? It's the new graphics that don't line up. (Not a great excuse, but the graphics are here to stay at least for a little while.)
The lack of a "refresh" option has been a problem with iCloud for years. Back in the iOS 8/9 days, I'd write in Pages on an iPad and then try to open the document on a Mac or the Pages web app. Pages itself was (and is) pretty nice, but iCloud sync was constantly broken. Things didn't appear when I needed them to.
Some designers say that refresh buttons shouldn't exist because the interface should always reflect the current state of reality. They're right, but until the day we get 100% bug-free bidirectional sync with perfect conflict resolution that instantly polls the network whenever it reconnects, refresh buttons are a necessary evil.
The only languages that eliminate logic bugs are formally verified ones, as the article points out. (And even then, your program is only as correct as your specification.) Ordinary Rust code is not formally verified. Anyone who claims Rust eliminates errors is either very naive or lying.
Type-safe Rust code is free from certain classes of errors. But that goes out the window the moment you parse input from the outside, because Rust types can enforce invariants (i.e. internal consistency), but input has no invariants. Rust doesn't ban you from crashing the program if you see input that violates an invariant. I don't know of any mainstream language that forbids crashing the program. (Maybe something like Ada? Not sure.)
I don't understand why you bemoan that Rust hasn't solved this problem, because it seems nigh unsolvable.