If somebody learns enough about organic synthesis to make cocaine after doing a lot of reading about it in a library (or two), the library isn't on the hook just because they had books people can read to synthesize knowledge.

If a library handed out an actual recipe with very clear instructions on making cocaine that also forgets to note how dangerous step four is, then yea, the library is liable.

I don't have a problem if these AI companies are held liable. You can't make massive money spewing out slop and go "oh, not our fault" when somebody slips and falls hard on that exact same slop.

It's getting a bit tiring to hear yet another example of how poorly the ecosystem created around these systems are and yet any reasonable criticism is too often met with some random variation of "AI is inevitable, adapt or die". Nothing here is inevitable, not even close.

OpenAI, the company, did this. It's not an employee or AI gone rouge. They were testing the model to do a thing and they did such a poor job of creating a proper testing environment that they created a major security incident with a competing AI company.

And they want to you to pay them to mine all your data so that they, well, trust them, it'll will be super cool and totally worth whatever we will charge you.

Put it back in the box and let it bake. Go fast and break stuff has limits.

There was a push in some factories to adopt a "30/40" approach. Schedule people for 30 hours, pay them for 40. There will be no real change in actual productivity (it may improve a bit), you can run three solid shifts and idle the machines to save power.

The pilot programs worked a treat. Probably didn't survive management coming in and not realizing that the 30 hours part was the critical part, went back to normal hours with a brief overall gain that quickly faded.

The forty hour workweek was designed with a single income household in mind. We'd all be better with an actual 32 hour workweek and strict overtime enforcement.

I look forward to the comments in which that is deemed lazy and how working 60 hours to mainly grow the wealth of the ownership class is the only proper American work ethic.

I mean, I sure enjoyed having to clean up the messes that overworked, slave to the grind programmers with delusions that stock options would be their way out made. Well, I didn't actually. Paid okay, however.

Not just DAWs, but all the VSTs out there are pretty much Mac or Windows. Sure, there are projects that support them on Linux, but the reality is so hit and miss that it is just not worth it. There are a ton of weird little programs people out there might use and like that keep people on Windows. If you bought a Cricut for random craft projects, you are pretty much toast on Linux.

As people note, there are lots of options for tasks on Linux, but for many things, people don't want a choice. Few people overall want choices for exploring the file system via a GUI or how to install a program and so on. They just want to cut out some craft paper and get on with things.

For Linux to get even close to how popular Windows or MacOS are, it would give up all the things that power users like about it. I don't know why some many Linux users are seeking this "year of the desktop" when it is a perfect example of being wary of what you wish for.

I fully expect this administration to just be an extension of Trump's ego at this point. Yes, going after companies that he disagrees with is a major issue (see how the FCC pressured major networks and how quickly some of them folded).

I can't care about Anthropic getting in a spat with Trump. Anthropic were the ones that started this "our model is so good at finding security exploits, it's too dangerous" marketing campaign and people are surprised that the current administration took their word for it and banned even their "safe" version after a simple jailbreak was found? There is a good argument that there are other models that are about as good as Anthropic's models when it comes to finding potential security flaws, but none of them when around saying "it was too dangerous to release to the general public".

They made their bed and somebody came along and lit it on fire. Seems par for the course these days.

I mean an AI would never look at a request to by a pair of wireless headphones and order a pair of wireless headphones. And by pair, you mean two orders of the same wireless headphones that were less than 150.00 each, correct?

This will be fun.

You have to interact with business and other non-government entities. Requiring they use ODT is a complete non starter, so using OOXML by default is the practical choice.

And yes, that means how Microsoft uses OOXML. The ISO standard is a starting point, but really, you have to also refer to the https://github.com/dotnet/Open-XML-SDK as well. Yes, this is a .Net library, but it works on .Net Core just fine.

Not sure that giving money to a third party to make your core product experience worse is the best of ideas. Google has the cash to see how it turns out, no question about it.

I am also very skeptical that there is any path to make the costs for LLM-based search the same as traditional search. I don't think Google can charge much more for advertising services either.

Of course, my bias is that people want predictable results when interacting with a search engine. Maybe I am wrong about that.

This idea of organizations replacing Office and Microsoft comes up over and over again, but in most cases, it just doesn't pan out. The alternatives are 80% there, but that last 20% is the crux of the issue. And the alternatives have been 80% there for decades now. As good as the political reasons are, people get annoyed when you replace things that work well enough and have for quite sometime with something that is in the uncanny valley of office software. It looks similar, it often acts similar, but it really isn't. Remember, I am taking about office workers. People that don't know Slashdot exists. People that have invested a lot in Word and Excel and Outlook (like it or not).

And there's not of a lot of savings there. Per person, you are looking at 60 to 100 dollars a month at the most. Yes, that adds up, but it is easy to lose a few hours a month in productivity dealing with the changes and that savings disappears. Also, you are contracting with Amazon, Azure, Google or (eep) Oracle for storage anyway.

If the EU really wants the US and US companies to pay attention, start building a EU wide government cloud and build alternative infrastructure around that.

I think there is increasing internal pressure inside Microsoft to leave Windows (and to an lesser extent) Office alone when it comes to AI. The pushback is growing from the userbase who just want to get things done.

Of course, AI bubble feeding Azure is also a thing. So, probably a good thing that somebody at Microsoft said "Android seems like a better bet for this" and the Windows group just nodded in agreement.

It is sad how many people don't realize how workable and reasonable this solution is. Sure, it requires one accept that most people will get paid better with a small precent making less money and that unions can be an overall positive contribution to the longevity and profitability of a company.

I mean, RAM prices went through the roof, union workers held strong and got big bonuses for it. Is Samsung going on out of business? No.

In the US, people have been convinced that they only thing companies should actually invest in is investors (the market). It's a theory that hasn't panned out. We've given it 50+ years; I think it is time to give more equitable distribution of corporate gains between workers, the community (and governments) and the market a try. Like what was done in the past. Sure, it makes to much harder to be a multi-billionaire, but nothing is perfect.

People have very legitimate reasons to reject how AI is currently being pushed into society. It doesn't just assist, it replaces. It makes people less skilled and more dependent on something that has enormous costs for questionable benefit.

Sign up or else you will be 10x as powerful as your peers. It's selling the illusion of expertise to those that don't want to expend the effort and preys on your fears of being left behind economically and socially. There is nothing utopian about it right now and there is nothing wrong with saying this is not wanted, needed nor tolerated.

Is there a future in which these are reasonable tools? Perhaps. But until, then, boo away.

Sadly, there are very few companies that could maintain a service like GitHub at the scale of usage that GitHub currently that and have it make sense in terms of a business. They got caught with not having enough capacity and giving out way too much compute in terms of Copilot because they got caught up in the AI bubble, and that's completely on them.

But other alternatives will have their own issues if they get a large bump in users. When you have that many people using a service, things get weird. Simple things are never simple, thing will breaks, people will try to break in more.

And AI is just adding a ton of slop that makes things worse from more and more angles. Glad Microsoft decided that all this Copilot stuff needs to start paying it's way. If it can't and it dies, fine by me.

Is that the NPU doesn't have enough TOPS for Copilot local features... And nobody really cares Microsoft is secretly fine with it, because the internal backlash to Copilot is growing. We have heard public rumblings about the Windows 11 backslide and that is just the tip of the iceberg. Too bad upper management is out to lunch and too high on AI slop.

I's easy to make Windows 11 suck less. Most of it is just undoing crap and a reasonable recommitment to actual native applications.

Compatibility is still an issue, because the amount of just weird that people run on Windows boxes, you can't keep up. On macOS, almost everything is ported over. Rosetta is just "give a shot if you really need something that old" safety net.

Panther Lake really closed the gap on battery life. I've seen some real life reviews and it remarkable how much better it gotten on the Intel side. Of course, there is still a gap to ARM (both Qualcomm and Apple Silicon), but it's not large enough to put up with the other hassles. Right now, the whole Windows on ARM thing is keeping Intel honest and in all fairness, that does seem to be working.

From a development standpoint, what a absolute nightmare. Cross-platform development is so hit and miss. Will a library support Windows/MSYS? It does, great. Try to build it for ARM64 windows target and more times than not, it just implodes. Heck, even ARM64-linux cross compilation isn't a given. The toolchain is much further along for sure, but it still a hassle.

And a salute for all those keeping macOS support going for those libraries out there. It's thankless stuff.