I'm not talking about what would be best for society, I'm talking about what would be best for Sony.

It does seem that way, since they probably would have expired otherwise, since they weren't able to execute. And neither can Intel. But since TSMC was able to advance process technology, we know it was possible, and therefore it would have been even better for AMD if they had been able to do that instead of divesting.

Widely? No. Divergent? Yes. Run the same query multiple times and you'll get different answers. The more input you give, the more subtle the differences are likely to be, but there's always a chance it will go completely off its nut and follow some line of "reasoning" that goes someplace utterly unrelated to what you asked for. This is less likely in special-purpose models, but never impossible.

But this is exactly where AI can actually be useful — so long as you don't depend on it to identify all of the design issues. It can review mountains of information and find things in it you might not have caught. You just can't assume that it will catch everything it "should" — that is, what it should be capable of catching is not necessarily what you are capable of catching.

Meanwhile, that periodic report has requirements, so you're going to have to ask it to write the report, then to review it to see if it met the requirements, then to review its own statements about whether it successfully met each requirement, because it won't do any of that sanity checking without being told to do so.

That's only even half of the stupid. There's no sense in a refrigerator even being able to be set to a temperature above about 70 F, as there's really no scenario in which you'd want to do that. You would just turn it off. Why on earth do they even allow these insane input values?

Is it just me or is it sick when Wikipedia doesn't even mention the singular thing people notice most about her, ostensibly in the name of political correctness? There's a way to write that section without violating their standards, but it isn't even there.

Cell was fabricated by IBM, so... No.

"because SSL makes it impossible to generate static content."

What SSL/TLS prevents is a caching proxy without doing any work. (Even that can be done, but it takes more steps.) You can deliver static content with encryption. If you are doing a lot of that, there are accelerators.

Have you seen this Supreme Court in action?

Both of these approaches have been used. Some vehicles have both things. For example a Sprinter T1N always has glow plugs, but a block heater is an option, but an Espar heater is actually more common! And when equipped, it comes with a timer, not just a knob controller like you would get with the cheapest install kit.

More hardcoded passwords and the like in the renewal processes where people have been doing them manually and don't know how to automate them correctly. These people should be outsourcing this...

You can't. You won't. You will outsource all of that to your web provider. Mine handles renewing my cert for me automatically. I could provide my own, but I don't have to. Instead, they handle getting it... from Let's Encrypt. I will not even notice this change, because they will continue to do so.

I have done my own self-signed certs and whatnot, I have installed certs from Verisign professionally yada yada, but there's no reason for me to do anything other than this for my vanity site. But there's really no good reason for most companies which can't afford to have an IT staff to do anything else either. They are not dependent on their website to do business.

I see them ditching their consumer line, since it's too hard to differentiate there now, and focusing on pro equipment. Their camera line is very successful.

It can't, though, and for two reasons. One, a person is required to check off on the data. Two, LLMs don't faithfully represent data. They ingest it as tokens and then what they crap out is influenced by random numbers produced internally as part of the token selection process. That means a person cannot certify that the data was faithfully reproduced without checking literally every value, because that's simply not how LLMs work.

You can potentially address this with a hybrid approach, where the LLM uses a tool to collect and manipulate the data, but this is shifting the problem to whether the LLM used the tool correctly. It might be more obvious when it doesn't, though, so it might still be a win.

What LLMs seem best at when it comes to text processing, and they are still not infallible, is summarization. But you can't trust its negatives, so it's only useful for identifying other possible positives. That's not worthless, it can help you catch things, but you can't trust it to do the parts you'd most want it to do.

I rarely have to adjust any options for a game to play correctly any more. Usually the internal benchmark does an acceptable job of choosing settings. Of late even pretty rinky-dink games have even been correctly figuring out that I want vsync (I have a 60 fps display, of course I do) which is nice, but they also seem to correctly pick the performance settings. I imagine that this is something that Unity and UE have finally figured out for them.

This has a really obvious answer which has been thoroughly explored here but I would like to explore why that answer is bad even though it's true, and pointing out it's true is not bad.

First some postulates:
Government should not be allowed to contract anyone to manage our data.
Government should not hire contractors to work on systems which will be used to manage our data.
Government should never be sending our data out to anyone else or allowing it to be stored on systems not under their control.

Obviously I realize that this is status quo stuff, it's more common than not in fact. Only legacy and defense systems are still held and run by feds and states, everything else is generally outsourced to a contractor (Accenture, Deloitte...) and is also probably running in the cloud — specifically on AWS, Azure, or Oracle.

But these things should never happen specifically because they are each individually poisonous to privacy, let alone all three together. All of this software should be developed in-house like it used to be. Government outsources software development and management so that it can do things it's not allowed to do otherwise, not because it's the only way to do it. It allows them, perhaps most importantly, to hand our data to others. Huge amounts of our PII is being processed not just on cloud instances provided by these players, but actually through their services (e.g. Amazon Transcribe) which are AI-based* and also mean that they actually process all of the data on our behalf and are in a unique and ideal position to harvest it.

* Amazon Transcribe used to support phonetic substitutions but now it doesn't, which I take to mean that they've just handed the whole job purely to a LLM which does it in one stage... you hope.