More importantly: Capitalism is an ECONOMY and market system. It is NOT a blueprint for a society. You can run your commerce and trade as capitalism, when you run your SOCIETY along capitalism principles you end up... essentially with the USA.

This is the part that is constantly forgotten. As a society, we have values that are not represented well within capitalism. But for some reason, we dumb shits think that we can treat everything as a market and apply capitalism to it and that will magically solve problems. But in education, just as one random example, the goal of it all is educated adults as output. It is not maximizing profit. Same for the prison system, the healthcare system and two dozen others.

It's greed, pure and simple.

Making a good product is possible. KEEPING making good products for decades is hard. Even more importantly: You will have hits and misses. Which, for a quarterly-result-bonus oriented manager is a no-no. Subscription models mean plannable revenue streams. Then all you need to do is negotiate your bonus package so that the already existing subscriptions will provide and you're home free and can already order your 2nd yacht.

Obviously the switch from "loss leader on a scale the capital markets can barely absorb" to "losing money" is going to sting; but I'm curious if we'll see sneakier knock-on effects.

So long as they were losing money hand over fist the vendor does want to throw enough tokens at you to make you feel like you are having a good time; but as few as are required to do that since they lose money on every one. If they were breaking even or turning a profit the incentive would be to sneak as much spend and upsell in as possible; and it's well known that the verbosity/cost of LLM chatter is hard to predict; harder if there are multiple models and other complications being switched around in the background.

What sort of exciting little tricks will we see from vendors who actually make more if you use more?

Which is, of course, nonsense. Nothing stops you from making a copy of encrypted data. What sucks is that for a backup, you likely can't do incremental.

In your case it might simply be that there was not enough of this kind of tasks in the training data.

I agree in general, but not with this strong phrasing. I've let AI build a good amount of non-trivial code. But my consistent experience is that it works best when guided by an experienced coder who can correct it, and when implementing well-known algorithms rather than coming up with novel solutions.

Example: I let it write up a quadtree implementation in a language for which there was no ready solution online. It took 2-3 correcting prompts to get a good result. I could've done it myself but it would've likely taken a few hours to get it all right instead of the half or so hour it took with AI. The important part for me was that there's nothing unknown in how to implement a quadtree. All the AI needs to do is take the 100s of existing implementations and translate them into a different language.

True but too simplified. The Luddites had an entirely different motivation: The fact that factories now employed women and children at very low rates meant that the men lost their status in the family as bread winners and head of household. That was a major social disruption, which we don't have with AI.

I'd compare it more to teamsters or wagoners when cars became common. Your job is threatened by a different way of doing the same thing, a way to which your skills don't cleanly transition. Some choose to pick up the new tech, some want the old ways to persist.

In the end, coachmen became chauffeurs, because rich people prefer to be driven around oder driving themselves, no matter if it's a horse or an engine doing the pulling. But much fewer teamsters and wagoners became truck drivers.

BOTH of them? :-)

Everyone knows that 10x engineers validate inputs and carefully manage dependencies; so clearly there's no risk of unexpected behavior.

It seems worth noting that one of the items in Wyden's rather pointed inquiry is the fact that the feasibility of doing this is known to have been demonstrated for the DoD by outside people familiar with it at least as early as 2016; so while this is the first confirmed case of adversarial use it's the outcome of at least a decade of just ignoring the problem; and a significantly longer period of failing to reasonably anticipate the problem. It's not like there's No Such Agency you could ask about "how could you spy on someone with the internet even?" if you wanted to know how well or poorly readily available information matched a nation state signals intelligence apparatus.

Purely as a matter of cellphones being expensive and somewhat tepidly capable in the before times I assume that there was a period within living memory when merely telling people not to Gordon Gekko on their DynaTAC where the russians can hear you was good enough; but that would have clearly and rapidly been getting less true for at least a quarter century.

The whole 'responsible disclosure' preaching and the not-terribly-subtle threats seem particularly bad given that there's an entire industry of actively more dangerous people who are not only treated as legal but actively courted by state agents and cops(and often even less savory customers, though they tend to be cagey about those); the ones who actively seek to keep vulnerabilities quiet so that they can continue to sell exploit tools and services based on them. Throwing zero days on github isn't ideal vs. getting them fixed; but it gets them fixed faster than if Cellebrite wants to hang on to a bitlocker bypass or Trenchant, and L3Harris Technologies Company, wants to keep selling 'network investigative techniques' that can bypass default windows defender configurations or whatever the situation is.

From the outside it's hard to know whether MS actually mistreated the researcher badly enough to justify their displeasure(the consensus appears to be that MSRC was never the best to deal with and has actively gone downhill; but this person's position seems significantly angrier than average) or whether they are perhaps wound a little tight; but implying that their legal status is the same as people actively running attacks against user systems is blatantly false and totally ignores the class of researchers who do actively run attacks while being treated as respectable.

It's a particularly bad look when at least Facebook got into a public legal fight with the NSO group over their nerd-merc work against their users; not like that actually solved the problem of attacks on cellphones; but it was an all-too-rare case of industry pushing back against the 'respectable' arms dealers; and not one that MS has an analog to.

The exact scenario we warned about when the discussions about this "responsible disclosure" nonsense started. Someone needs a reminder that letting you know your software sucks is a courtesy, not something you can demand.

We'll all pretend that nuclear-tipped VCs will remain cool-headed and cooperative when the AI debts start to go sour.

I'm not an expert on rebuilding massively exploded launch infrastructure; but I have a suspicion that a 2026-2029 plan is now going to involve less Blue Origin than previously believed.

Yeah, their slaughtering of possibly tens of thousands of protesters was clearly a sign of upcoming concessions.

Dictators lose when they make concessions. They stay in power when they double down. That's the hard lesson of a hundred years or so of dipshits becoming big boss by military coup or revolution. Those who put absolutely every penny into propaganda and oppression tend to hang on to power the longest.

And given what the IRGC and the regime have done to the Iranian people and how much they're loved in the rest of the world, staying in power is literally a life-or-death matter for them. The day the regime falls, we'll see all the Ayatollahs and minions hanging from trees.