Comment Re:Kernel or userspace? (Score 1) 23
Why is this still a thing? FFS. A stack overflow? Why. Is. This. Still. A. THING?!
Could we cut it out with the C# already? Computers are powerful enough that you no longer need that sort of latitude.
Why is this still a thing? FFS. A stack overflow? Why. Is. This. Still. A. THING?!
Could we cut it out with the C# already? Computers are powerful enough that you no longer need that sort of latitude.
I just got mod points, and looked in on responses, and I couldn't agree more too.
I almost want to delete the OP and mod that one up.
The thing here is that you're burning coal for the purpose of not having to burn coal again. You'd have to have an attention span of a gnat to not see that far ahead.
We can't go straight to zero emissions guys. There's gonna be a transition. This feels like clickbait, lacks long-term thinking, and is not a reasonable argument.
What are you on about? I've used Thunderbird since version 2. There was some update that broke extensions, but they literally haven't broken in years.
If this is an exploit in the anti-cheat driver, that's a far more serious matter.
I hope it's a userspace hack.
AGI is not jargon. It's to distinguish an important milestone from the original term, "AI," that has become a marketing gimmick as opposed to the technical term.
It's kind of how European scholars have to distinguish liberalism from classical liberalism because US marketers ambiguated the original term.
None of this is AI, and LLM natural language prompting is a loser of an application.
There are lots of better uses for large scale pattern recognition, and once we get over how human a LLM actually sounds as a gimmick, we'll find actual uses for it, like detecting heart disease. It turns out patterned natural language output is fatally flawed in my opinion, but pattern recognition itself is game-changing.
That it actually fooled some people into believing it was "sentient" is the reason we're faffing about with this dead end tech niche that has no real appeal to anyone but fans of Microsoft Bob.
And you thought pop rocks and coke was a killer combination.
Imagine what would happen if you ate a laser, drank some diet coke, and had a mento. Pew Pew Pew!
Is China, an ancient society and the second most populous country on earth, in the same ballpark as the United States when it comes to the future of the human race? Could they even do it better? Aren't they some third-world country? When did this happen? Why are all of our scientists leaving, for that matter?
Humanity is advancing and it's not the USA. Stop the presses.
But... ZOMG! Is this actually a problem?
Only if you're an American. Honestly, I've got my popcorn out, because I'm not sure "beating" other nations to something closer to AGI is going to work out well for the country that invents it. It's 50/50 if it creates a new global hegemony, or wipes their culture off the face of the earth, imo. It's a race I don't want to win. At least not for nationalist purposes. I think application of what we're calling AI for national dominance might just backfire on those who use it.
TFS for this:
Crashes are caused by physics.
Most of those Americans with a sixth grade reading level choose not to read anyway. It's a fair choice. (It's also an overstatement. When I write marketing copy, it's to an eighth grade audience if I want to hit the mainstream.)
But to quote Egon in Ghostbusters, "Print is dead." Really. It is. It just took a few decades.
I had a discussion about it with my tenured professor wife, who thinks that civil society is collapsing because people do not have high-value information due to lack of reading. She also blames a lack of basic literacy for many other historical problems. If only those people read, they wouldn't have had that problem.
Frankly, all of us smarty pantses need to suck it up and get on YouTube or Tik Tok if we want a high information society. There are tons of channels on which to disseminate, but it requires acceptance that most Americans do not read. Neither did the ancient Greeks, for that matter. They had an oral tradition. Aristotle literally railed against the written word's detrimental effect on people's memory. It's possible to have a high information society without the written word. Other traditions and methods of dissemination work just as well.
So whatever we're doing now would work if some people would get their snobbish noses out of their books, exercise restraint instead of deriding choices not to read, and their get their butts on media people actually use.
I'm sick of the "people don't know how to read, and should read more" argument. It's elitism.
So, they're cramming a bunch of X-Box stuff into the Windows codebase?
They do know that gaming s__t is inherently unstable right? Just ask anyone writing video card drivers.
What will this do to Windows stability? Oh, right. Business customers will be allowed to get a version without the entire thing.
So only consumers will be screwed. Yup. Par for the course. They may as well fork the operating system at this point. "Windows 11 for Business" and "Windows Suck-It Beta Tester Edition."
You can't make this up. I started a novel in 2002 where the Internet went nuts and we had connected every possible thing to it, making us have to smash all our clocks, cameras, watches, appliances, etc and go low-tech.
Air Fryers was not on my BINGO card.
Hey, Microsoft Office also just stealth turned on Copilot for me! When it launched, it asserted itself in the ribbon in Word. I turned the icon off. No interest. I am an expert writer.
Apparently a lot of people did the same, because...
Today, there's a little gray "Draft with Copilot" icon that appeared in the margin as I was typing. Right clicking on it gave no options to get rid of it. I went on a what I thought would be a fishing expedition in options. It was easy to find. There was a "Copilot" tab. It was turned on. Bonus points to Microsoft for not burying the setting, or not having one like they started out with the Edge sidebar. Boo Microsoft for making it less convenient than a right-click.
Yeah. Adoption. What's unprecedented here is the aggressive forcing this on the user repeatedly. I've had to turn this off in Office twice now. I wonder how many people just ignore it and don't turn it off? Are they part of the "unprecedented adoption?"
In light of the "adoption rate," I want to make it clear that Google has recently hijacked anyone's phone using their Assistant product and switched it to Gemni.
It can be gotten rid of, and you can return to using Assistant. I'm not going to post a tutorial. I asked Gemni, which I found to be deliciously ironic. It gave a deceptive answer on the first prompt, so I refined my prompt and it finally said "Good News! You can go back to using Assistant," and then proceeded to give me an inaccurate process, but enough information to figure out the actual process.
Gemni usage collection is carefully collected and then buried a screen's worth down on your "Other Activity" page under "Gemni Apps Activity." They'll advertise it's potential, but boy are they obfuscating its existence.
I think I was successful in getting it off my phone because that page shows no activity now. The day it went live it showed all my Assistant stuff as Gemni interactions.
I found out because Gemni accidentally activated in my bathroom, due to an unknown prompt that was not "Hey Google," spoke in a male voice and asked me what I wanted. Oops. Guess they should have covered their tracks better.
How much you wanna bet this nonsense is baked into "adoption" figures? There's a difference between adopting a product and being silently hijacked.
There's a difference between "adoption" and people engaging in fuck around/find out exercises.
I'd like to see where the data distinguishes between the two.
I've done a lot of FAFO with all the LLM models. The only remarkable thing I've noticed is the dishonest way it presents itself as an AGI, and its simultaneous willingness to tell you it is not an AGI (if you know how to ask), while still being stubbornly unwilling to present itself as a tool rather than a [just-add-random-seed] "person."
I'll take ChatGPT. You should have seen how hard it was to program ChatGPT to remind me every third or fourth message that it is not AGI. I got there, but it took a lot of prompting. I addressed obvious boilerplate as an "error message." It insisted it was not. Hell, it even insisted it was generating the boilerplate, when the resulting message was word identical, which is literally impossible for a functional LLM. I asked it about syntax, and variables, and logical operators and it steadfastly refused to accept basic computer science prompts asking how to work it. It wants stupid people to believe more than it wants to be a useful tool.
Which tells me that the devs are taking their cues from marketing. Maybe with a wide eyed sense of "ooh it's cool!" guiding them towards compliance. It simply does not behave in a way that I find useful.
My conclusion is that the current purpose of ChatGPT, and others, is to deceive the potential market as a stunt. Whether or not it exceeds that developer mandate is not going to predicted by a venture capitalist with self-interested bias. This article is hype.
GIGO.
And that doesn't even go into the fact that it is literally being forced on people, which is definitely not adoption. It's a hijack. My entire phone switched to Gemni without telling me, and when I found out because Google was sending me aggressive Gemni marketing emails about all the "new things my phone can do," it took a lot of hacking and a half hour to rip it out of there.
By the way, it carefully warned me that Gemni is inaccurate, "especially in regards to information about people." (paraphrase)
Great. They replaced an Assistant product that knew its limitations, and was willing to say "I didn't understand that," with a marketing stunt that ignores them and always has "the" answer, even if it is a complete fabrication.
This is not progress. It's a scam.
"Falling in love makes smoking pot all day look like the ultimate in restraint." -- Dave Sim, author of Cerebrus.