In that case, how does the week's training help at all?
Maybe it's helpful just for you to understand the way the military is organised: if you are conscripted you should report to this base, you'll sleep here, your commanding officer will be someone from this branch of the armed forces, you'll be in a group of X people sharing Y shifts, etc.
I think maybe there's another step too - breaking the design up into small enough peices that the LLM can follow it, and you can understand the output.
So do all the hard work yourself and let the AI do some of the typing, that you’ll have to spend extra time reviewing closely in case its RNG factor made it change an important detail. And with all the extra up front design, planning, instructions, and context you need to provide to the LLM I’m not sure I’m saving on typing. A lot of people recommend going meta and having LLMs generate a good prompt and sequence of steps to follow, but I’ve only seen that kinda sorta work for the most trivial tasks.
Unless you're doing something fabulously unique (at which point I'm jealous you get to work on such a thing), they're pretty good at cribbing the design of things if it's something that's been well documented online (canonically, a CRUD SaaS app, with minor UI modification to support your chosen niche).
And if you are doing something fabulously unique, the LLM can still write all the code around it, likely help with many of the components, give you at least a first pass at tests, and enable rapid, meaningful refactors after each feature PR.
There are some economics of scale that work best at the country level.
Even with the EU single market, mobile phone operations almost always follow country borders. You'll get a different set of providers in Germany than you'll get one km away on the other side of the Rhine in France. Even though some of them may have the same name or the same ultimate owner or both, and even though you can roam on the other side of the border, you'll have a contract with a different entity, and different people will build and maintain the networking equipment.
Conversely, in the US, the major carriers all have nationwide coverage.
It's an order to form a standing army, not an order to deploy troops anywhere in particular. The article describes a recruitment and training drive that sounds like it should take 3-12 months to complete.
If you do believe (as your post suggests) that an armed response to crowd control might be appropriate in the event of another Vietnam or another BLM, surely the right time to start preparing for that is now?
I do think you're right to be concerned about militarisation, but "there isn't enough civil unrest to justify this right now" is hardly the right objection.
The point I was making about "enough civil unrest" is that the administration is trying to use it as an excuse to deploy the military in ways they want. It's cover for their true purposes.
I don't believe they should be deployed hardly ever, if at all, and there is a good reason they are prohibited from being used as domestic law enforcement.
Do you understand the concept and purpose of the military industrial complex?
It is, in no small part, a capitalist enterprise. War is a means of projecting imperialist power, but also justifying military contracts and expenditure. It is a self-perpetuating, self-justifying machine that demands constant growth, so we must do that. We can't not do that.
And since the current regime (and half the country) considers the domestic population to be as much an adversary as any foreign one, we're going to reap the home soil. This is just American imperialism turning on itself, as imperialism inevitably must.
I thought the writing was banal but fine. The stars of the show are the images, not the text. Assuming it's largely LLM written, this seems like a good use of that technology.
The text adds some pieces of information you wouldn't get from the images alone: they were painted on flour sacks, used at mobile cinemas, now exhibited at galleries in the West, etc. And it provides citations and artists' names for those who want to learn more.
The art criticism is unsophisticated, the images don't completely match the descriptions, and some of the facts might well be hallucinated or at least taken out of context. But you got that with traditional media and human writers/editors too.
For what it's worth, I'd guess there is a real author, whose command of the English language is worse than ChatGPT, though his personality is more interesting, and who asked the LLM to rewrite his work in the right style for the website.
> For what it's worth, I'd guess there is a real author, whose command of the English language is worse than ChatGPT, and who asked the LLM to rewrite his work in the right style for the website.
Sure. But if the author doesn't notice the nonsense that the LLM is introducing, it harms as well as helping.
"Primed for colour" is a strangely uninteresting thing to be saying about the sacks. If this requires any non-trivial effort, it would make more sense to describe the process. If the author actually wanted to talk about that, chances are the LLM removed useful information.
And putting aside that "These weren’t just any flour sacks either — they were durable, easy to roll up, and ready for reuse." is three "classic LLM tropes" in a row ("not just any"; a gratuitous emdash where any dash at all only becomes necessary because of that introduction; an ascending tricolon), it's just a bizarre thing to say. First off, if the sacks were sewn together to make a larger banner, then it doesn't make sense to talk about rolling up the individual sacks. Second, the phrasing suggests something exceptional, but these are all totally ordinary and trivial properties of pretty much any sort of flour sack. Many different materials are used, but all of them would be "easy to roll up" when empty, and making them durable and reusable is just common sense in that environment. The artists were clearly just using a fairly obvious material they had at hand, so this sudden bit of marketing-speak is entirely out of place. Third, the features highlighted all have to do with the sacks, but not with either each other nor the banners. In particular, a sack being "ready for reuse" is ready for reuse as a sack, not for its material being repurposed for something completely different (we typically call that "recycling", not "reuse").
The bit about "the designs" may well even be true, but it's a complete non-sequitur here, a point that doesn't really merit deeper explanation.
The writing isn't just "banal" but nonsensical in context, veering off into free-association. There's more potentially being hallucinated here than just the "facts". Never mind the accuracy or truth of what's written; this sort of thing makes it hard to accept that the prose even reflects the author's intent.
> The text adds some pieces of information you wouldn't get from the images alone
But this is exactly where being AI-written bothers me! I don't really mind the style (the LLMs have learned to write a particular way because 1. people write that way and 2. other people like it) and I don't have the "boooo stochastic parrot plagiarism machines booooooooo" sense of disgust at AI that some people have, but I do know that when LLMs write things those things are ... not always true.
(Of course when people write things they aren't always true either, but the LLMs get things wrong more than humans do.)
Which means that when the article tells me something interesting -- flour sacks! mobile cinemas! exhibited in galleries! -- I can't trust it. And that, for me, is the main damage that outsourcing your writing to an LLM does: it destroys trust.
Still the ones who got in would be with more other students who opted in, the ones who didn't get in would be with more students who didn't opt-in to the lottery.
You would need to have a second group of those who lost the lottery and were all put into the same non-Montessori school with no others who didn't opt-in maybe.
Yeah but even then teachers that opt in to train in Montessori might just be better teachers, and converting a whole school system to Montessori, training everyone, might not have as good results.
A bit of both. He definitely did criminal things, but they look worse because Gaddafi was such a politically unpopular ("terrorist") leader in the west. If he'd got the same funding from the Obama regime, surely he would never have gone to prison.
There are so many different political perspectives that would inspire someone to use the word "regime" to talk about Obama's presidency of the US, that I'm genuinely not sure which one the parent commenter is likely coming from. It's not a dog whistle it's a whistle for every type of animal.
I mean, I don't think we need to put air quotes around "terrorist" for Gaddafi. This was a ruler who was happy to bring down Western passenger jets and put bombs in night clubs.
Options pricing is reasonably competitive. Even a gambly thing like a Tesla zero day option has a spread of 1-2%, so someone trading it at random loses 0.5-1% per trade. And Robinhood is a brokerage, not an options market maker, so it doesn't capture all of that 0.5-1%.
You'd have to read Robinhood's financials to see what they mean by gross margin. Possibly it means if a customer deposits $1,000 and trades options, the customer eventually on average loses $900 of it? Even that seems too much TBH.
It is not a good idea for retail investors to get heavily involved in zero-sum derivatives trading against much more sophisticated algorithmic trading models.
It's just not for you to judge. Those with sophisticated models usually have a lot more capital to manage. Also, it's not a zero sum game because alignment with the underlying's price drives it. Long term share holders are the foundation that makes it not be zero sum. As a retail account grows, its approach too can become slightly sophisticated over time.
It's not a good idea for punters to go to the casino and bet it all on black. Some percentage of the population is always going to be degenerate gamblers. We can try to reduce the harm a bit but ultimately this is just a reality we need to accept.
One of the weird things for computer people about the Irish voting system is that it's non-deterministic! You can count the same ballots in a different order and get a different result (because it depends which votes you choose as "surplus" to redistribute).
In practice it doesn't seem to matter that much. The counters even out the first-level effects of this, so it only matters for votes that have been transferred more than once; it can be determined statistically that it changes the result only in a very small number of cases; and there are plenty of other weird threshold effects to care about instead. But it's one property you might expect of a fair voting system that Ireland doesn't give you.
Yeah. I think it's the best voting system in the world because I've yet to encounter one I think is better but you're right, it's far from perfect.
That said, surplus distribution tends to be the main flaw raised time & time again, & whenever improvements are discussed the general conclusion tends to be that the current distribution mechanism goes a very long way toward fair representation of the actual preference distribution. It's notable that the more computationally intensive alternatives to get "fairer" outcomes are pretty recent inventions & it's really hard to justify the effort given the tiny number of cases affected.
The engineers who wrote your browser already thought of this and made sure it wouldn't work.
In case anyone mocks you for this, though, it's not a stupid question at all: there have been 1-click and 0-click attacks with vectors barely more sophisticated than this. But I feel 100% confident that in 2025 no browser can be exploited just by copying a malicious string.
>But I feel 100% confident that in 2025 no browser can be exploited just by copying a malicious string.
that's a real far leap. Most OS have a shared clipboard, and a lot of them run processes that watch the thing for events. That attack surface is so large that 100% certainty is a very hard sell to me.
Just for the sake of arguement, say clipboard_manager.sh sees a malicious string copied from a site by the browser to the system clipboard that somehow poisons that process. clipboard_manager.sh then proceeds to exfiltrate browser data via the OS/fs rather than via the browser process at all, starts keylogging (trivial in most nix), and just for the sake of throwing gas on the fire it joins the local adversarial botnet and starts churnin captchas or coins or whatever.
Was the browser exploited? ehh. no -- but it most definitely facilitated the attack by which it became victimized. It feels like semantics at that point.
This is a good point and it completely fits serf's concern. So OK, I change my answer, it is reasonable to be concerned about exploits from just copying malicious content to the clipboard.
Maybe it's helpful just for you to understand the way the military is organised: if you are conscripted you should report to this base, you'll sleep here, your commanding officer will be someone from this branch of the armed forces, you'll be in a group of X people sharing Y shifts, etc.
reply