Anybody using this is already running Home Assistant for other purposes. The plug draws a watt (*drops monocle*) to adjust for ToU electricity usage. Simple goal, simple project. Do people like you suffocate if you aren't constantly bitching about something?

If you're a big business, you'll save money going with a M365 subscription over a perpetual license.

Why? Because M365 is licensed per user, while Office is licensed per-computer. M365 allows up to 5 PCs to use the same license (which may include a personal install on the user's home PC), while office perpetual is licensed just on that PC.

If you've got an employee with a desktop and a laptop, they're going to have to have 2 perpetual office licenses, or one subscription. Given the costs are roughly the same (if you buy a new version every 3 years or so) not having to have two licenses is cheaper.

In additional Microsoft is part of the gang of companies that can demand accounting of all the licenses in use. So you start having to produce receipts of purchasing as many licenses as computers. M365 being licensed per user makes this much easier to manage since you only need one per employee.

The tricky thing will be what it does to the talent pipeline. In practice a lot of the high-clue programmer supply is either trained or discovered by initially doing the somewhat lower-skilled stuff and gaining experience and working with more experienced people.

If companies think that they can replace their entry level programmers with bots they'll presumably do so; but if there are basically no entry level programming positions to be had it's unclear who will be gaining experience to become the more senior people who will be required in specific instances.

Who will buy all that sweet, sweet private info without the ability to sell it?/blockquote

Lots. Companies that deal with LEOs will love to get at this data - they can offer searches of all that genetic data to help find potential suspects in crimes.

Those companies don't need to sell the data, they just sell a service to scan the data for potential partial matches for a subscription fee.

How much longer will there even be an FTC? Not holding my breath for enforcement in the current environment.

Not long, wasn't Elon Musk being investigated by the FTC? He's been going through the agencies that have been investigating him or his companies, so I'm surprised there is an FTC to speak of right now.

but in the 1980's it stood for the villain of the Tron series, Master Control Program.

Of course, the Master Control Program is really just another term for Operating System these days. So tying the two together we end up with Windows and CoPilot....

Good news, they just divided the estimated amount of wealth by the estimated number of people in two scenarios.

Unfortunately, my tendency is to view it as bullshit, not because of the climate science, but because of the economic models, where mitigated climate change is depicted as a simple differential equation of reduced exponential growth and unmitigated climate change is rendered as the same thing but with a new factor in the form of supply chain disruption estimation, which is essentially derived from averaging 14 other published models that were designed to measure other economic impacts of climate change, but not exactly the ones this paper is assuming.

Basically, my objection is that supply chain disruptions are not exponential in nature, and there's an adaptive dimension to them such that their effects on the flow of economies is dampened by planning.

Finally, some sanity in the whole discussion of AI coding.

No, our software development jobs aren't going away any time soon. The typing part is just going to get a little easier.

There always was sanity. If you watch vibe coders at work, you can tell the tools are nowhere near where they're hyped at. Sure the early days they seemed cool enough where ChatGPT would get you a lot of code that seemed to work, but ask it more sophisticated problems and things break down really quickly. There are plenty of videos of people doing it.

Also, check out the App Store or the Google Play Store and you can see some categories of apps have exploded, because that's what the AI coding would generate. Things like "meditation" or "mindfulness" type apps have exploded in number.

Of course, the hype will still be there because we're in the AI hype cycle, but if you've seen the jobs asking for vibe coders, you'll be less than impressed. Most of those jobs are paying not much more than minimum wage.

And I think at best we'll see the destruction of a new job - the prompt engineer.

AI slop is still junk, regardless of if it's a photo, a picture, music, books, papers, or code. Basically the only thing AI does is provide the initial bulk, it still requires a lot of human input, modification, revision, and other things to turn it into something useful. And chances are if you tried it, it would take you longer than if you did it yourself. At best, it might provide inspiration.

That's not to say AI didn't benefit me - but its contributions had to be carefully monitored and in the end I didn't use the vast majority of what it provided. It did provide help with some troublesome phrasing, but that's it. The rest of it was just crap, slop and hallucinations and was unusable.

I'm glad to see that "vibe blue teaming" has also come of age in our AI-enabled future!

That said, this whole thing won't be properly silly until they get hacked by ignoring some totally banal dependency on a web facing host while they are busy frisking every unintentional radiator and suspected finite state machine in the building.

AMD is doing very nicely with their GPUs nowadays, because nVidia basically exited the gaming market. Sure, they have the top dog GPUs, but that's it - they aren't concentrating on selling those cards very much hence all the shortages. That's because gaming GPUs make up a tiny amount of nVidia's revenue so they're concentrating less on gaming, and more on AI and other things.

It's why they just price their top end GPU at $2000 and call it a day - they don't care if people buy it or not. If gamers want to pay for that extra FPS, nVidia will sell it to them.

Meanwhile AMD is seeing good sales of their high end card, which can't compete with nVidia's top end, but is very competitive at the price point for midrange cards of $500 and under. They are selling through every card they make because you just can't get any nVidia cards.

Meanwhile, Intel's Arc cards are good for the lower end are surprisingly good on the latest handhelds using them

I don't see Intel chomping at the midrange level, but the low to mid range will probably work for them in the budget gaming arena

Also some astonishingly bad cell phone modems.

Except that Intel sold that to Apple, who managed to turn them around into a surprisingly decent modem in the C1 chip on the iPhone 16E.

It's apparently quite competitive with the Qualcomm modem, except lacking the mmWave support to achieve the ultra-high-speed transfers.

Except for the experience of going with a group, I don't see a reason to go to a theater.

Says the person who lives in a house, likely in the suburbs.

Because a large TV, nice sound system, etc, are much less possible if you live in the city, so you're likely watching movies on a 50" TV with TV speakers or a soundbar turned on low.

Megaplexes might be hard to get to since they're almost always located out of the area, but smaller theatres are located in many urban centers offering people who prefer to walk or transit to get around the ability to get big screen and big sound.

So maybe it's dying in the US, but if your apartment's the size of a shoebox, they aren't going anywhere yet. And yes, there's always the option of moving, but then you often need to then acquire a car, with maintenance, insurance, etc, and then sit through long commutes often in bad traffic. So there are tradeoffs.

yeah imagine having all that money and still caring so much what randos think of you.

What a petty, petulant child. I bet he and the orange idiot get along well.

You forgot about Elon Musk, the world's richest man-baby. He bought a social network just so he could control what people were saying about him on it. Then he started buying elections because people stopped using his social network.

Been running MacOS and Linux at home, and Server 2016 via RDP at work for a few years now, and had the misfortune of setting up a Windows 11 laptop today.

All I can say is that Windows sucks horribly. I cannot believe just how unusable, cluttered and full of marketing crap it is.

I suspect that actually having a mechanism to do this(much less avoiding having that mechanism succumb to the obvious adversarial move of an ISP or similar declaring all domains as 'internal' when resolving them for users) would be deeply nontrivial; but it would be really nice if there were some way to cleanly distinguish between 'internal' and 'public' use cases for SSL certs.

Unless you are talking a toy-scale situation of telling a handful of admin endpoints about the thumbprints of a dozen self-signed but known-good certs; it's probably easier to go with at least some fairly weaksauce PKI for all the various appliance web admin interfaces, as a better alternative to PEAP/MSCHAPv2 on WPA2 RADIUS, for smartcards, etc. However, it's often infeasible or undesirable to leak internal host information for certificate transparency or automated CA verification purposes; and it's also a situation where the security risks of incorrect issuance, certificate lifespans, etc. are pretty much on you. If Eve and Mallory are issuing certs for your vmware infrastructure you are probably pretty screwed; but that's not really a public issue.

The 'public' cert case, by contrast, is one where there is strong justification for being intrusive and humorless because we all have to deal with the fact that using 'trusted' CAs is a messy hack that means that it's possible for the dumbest, most malicious, or most legally entangled CA out there to compromise the integrity of encrypted connections with any domain(for at least some number of users; whoever gets the malicious cert still needs to insert themselves into the conversation; you can't just sign your way into retroactive access to recorded session data; unless someone really screwed up); and it's also the one where the downsides of demanding transparency and obnoxious validation requirements are lowest because the affected domains are already public internet facing(or, at most, 'private' but interacting with large numbers of unmanaged/unaffiliated devices; as in the case of something like guest wifi captive portal pages, which may not be visible from a random spot on the internet but are not meaningfully 'internal' in terms of their clientele).

As noted; I suspect that what I'm asking for would somewhere between 'nontrivial' and 'intractable in principle' to deliver in a way that wouldn't be readily abusable to facilitate a lot of SSL snooping on what are basically 'public' cases; but it's still an ongoing problem that there's not a great way to be restrictive enough for the awful world of public certs; where any of at least several dozen random entities of questionable trustworthiness could potentially break the system for any domain if they act in error or malice without being detected; without being unjustifiably or impossibly restrictive for internal environments where you just need some garbage appliance web interfaces, some of which don't even pretend to support SCEP much less the new hotness mechanisms for public issuance based on proof of control, to show up correctly on a small number of systems IT owns and controls. (Or, of course, the super-cursed situation of needing 'private' garbage to be able to interact with 3rd party systems allowed onto an internal network for specific reasons; or with devices that don't support certificate management on the client end either because they suck(either out of apathy or because their vendor's security model is user-hostile and they don't want their ability to phone home to the mothership for DRM purposes interfered with or whatever).)