If it's true that xAI is really about answering deep scientific questions — AKA understanding what the hell is going on — then this is a nice contrast to Mark Zuckerberg's Threads, which was created (it seems) to compete head to head with Twitter.
I went into the comments over there, mainly to see if they were saying what I expected them to say — essentially: We hate Musk. And, yes, there it is, right on top of the "most liked" pile of comments: "Just what we need, AI that is explicitly fascist and white supremacist."
IN THE COMMENTS: Leland writes:
WaPo commenters are why ChatGPT does so well. Their comments seem like coherent sentences, but the words used don't create a coherent thought. What is explicitly fascist? Was Musk proposing having his AI collude with the government to control the population? That's fascism, and while it exists quite a bit these days; I don't think that is what Musk is proposing. Musk is proposing quite the opposite.
Instapundit quotes that comment in full and adds: "Nowadays anti-fascism is called 'fascism,' just as explicit racism is called 'anti-racism.'"
Posts
61 comments:
WaPo commenters are why ChatGPT does so well. Their comments seem like coherent sentences, but the words used don't create a coherent thought. What is explicitly fascist? Was Musk proposing having his AI collude with the government to control the population? That's fascism, and while it exists quite a bit these days; I don't think that is what Musk is proposing. Musk is proposing quite the opposite.
"And, yes, there it is, right on top of the "most liked" pile of comments"
So not your shocked face? Got it.
"During a live audio event on the social media platform, Musk’s team of all-male math, AI and engineering experts spoke about how they wanted to create an AI that was smarter than humans with the goal of benefiting society."
Or we humans could just make First Contact with the Vulcans.
That relationship was rocky for the first century, though.
AI can’t make accidental discoveries the way their human counterparts do.
Unless ‘accidental discoveries’ is another way of saying ‘Devine Providence’.
Which could mean that it has been divinely provided that we discover what we’re meant to discover by believing that WE are doing the discovering while maintaining to ourselves and each other that it was AI that did the discovering.
And the man said, The woman whom thou gavest to be with me, she gave me of the tree, and I did eat. - Genesis 3:12 KJV
Conclusion: AI is no guy.
It is truly frightening how effective democrat propaganda is with respect to their acolytes.
Leave it to the Post to worry about the lack of a token female on Musk’s AI development team.
Blah, blah, blah blah, blah, blah, blah, Musk's team of all-male, blah, blah blah, blah, blah, blah, blah, blah blah, blah, blah, blah.
Why is the Post wasting all this ink and these blahs? Only one thing matters in the whole story.
"Just what we need, AI that is explicitly fascist and white supremacist."
How can software be physically fit? Wouldn't that be hardware?
"All male."
The horror!
Radio preachers in the early 80s, one in particular, were warning about artificial intelligence, only this preacher called it "A One" because he'd never heard it pronounced. Anyway the warning has to do with the mark of the beast.
You'd get all male if you're into deep philosophical questions. Women don't have the necessary obsession for simplifying, quite the opposite. Women add complexity until feelings solve the resulting complex problem.
What will really convince me that AI is a thing to be reckoned with is if one of them solves a Millenium math problem like the Riemann Hypothesis or the Birch/Swinnerton-Dyer Conjecture.
Well at least the team was not 'all white'.
How do I know that?
If it was so, it would have been mentioned in the headline.
They can’t get lane keep assist to work right. Full autonomous driving is looking like a scam…but you do you.
Guy can get us to Mars and do AI, but since the day he bought Twitter, logging on to my timeline is no sure thing. "Welcome to Twitter," it often declares with a "Let's go" button that doesn't go anywhere. I've also been throttled down by at least two-thirds on a typical tweet. I guess I should be grateful for the opportunity to spend less time pretending that I'm not wasting time.
If I was a Post reporter — and hadn’t yet committed suicide in shame at my profession — I’d want to know know whether xAI will find it as easy to prevaricate and to reference invented “facts” that aren’t real the way that ChatGPT does.
We tried to put the face of a female in charge of an AI discussion earlier this week. What we got was Kamala doing Kamala. She headed a 'roundtable' filled with women from various ends of the social justice world, all oohing and aahing at Kamala's accomplishments. It was a highlight filled talk. I loved when Kamala explained that "AI is about machine learning" (then looks at notes to find out what she meant by that). It went down from there. Kamala noes best
I think Musk is on a better trail using the best possible engineers and scientists that he can to work on this. He will, once again, work at dragging humanity, kicking and screaming, into the future.
Let WaPo and it's readers check off boxes for skin color and vaginas, because, you know...humankind's future depends on counting skin color and vaginas.
Musk's team is all-male! That will make all the difference.
Math is hard, right WaPoo?
Now that a few different groups have breathed life into A.I., we can move out of the Godhead phase and into the agenda-building phase, the A.I. wars. We already have a few decidedly left-leaning-so-far-they're-horizontal versions that can't stand to even consider Donald Trump's name. So, now that it has been proven that these bots are all a product of their programming constraints, it's down to Musk to represent a different side, different aspects of machine intelligence. Why not one that simply pursues science? Is it a sin to exclude politics from machine intelligence? Some of the Gods among us think so.
I guess we can assume that we're safe from Republicans building an A.I. that solely, relentlessly points out deficiencies in Democrat policies and highlights these with citation after citation of proof, and then goes on to show all the various Democrat political hypocrisies of policy one day, and opposite policy 5 years later - complete with video clips. No danger of that - it might mean Republicans winning the smart way instead of accidentally, and we can't have that.
"Just what we need, AI that is explicitly fascist and white supremacist."
Oh, the illiteracy... I am literally shaking.
Good for Musk!!! I bet he looked for the top mathematicians and programmers.. and the best were males so he used males... no PC 'Affirmative Action' stupidity.
He wants answers to hard questions before he creates this AI... he does not want a 'Colossus The Forbin Project' or 'Terminator' outcome.
He is not kidding around. He wants the real deal.
Musk's team is all-male.
Eat your heart out, Anheuser-Busch!
I am still surprised by how much the left hates Musk just because he is for free speech. He isn't very conservative by historic standards. The left is bent on extinguishing all dissent.
"Blogger rehajm said...
They can’t get lane keep assist to work right. Full autonomous driving is looking like a scam…but you do you.
7/15/23, 8:35 AM"
And you have some sort of point? And I expect you will be the first to say you were wrong when it comes online, which it will, because Musk will force it to.
"yes, there it is, right on top of the "most liked" pile of comments: "Just what we need, AI that is explicitly fascist and white supremacist.""
So will sheer inane prog predictability easy for AI to crack and reproduce or, because it is so not-I, very hard?
Rhhardin said: "Women add complexity until feelings solve the resulting complex problem."
Huh. Especially the feelings solving problems part. More like creating problems.
Of course the team is all male. They are teaching the AI logic not emotional responses.
As long as they make an AI that doesn't make shit up, I don't care what gender/sex they are.
There's an intriguing hint in the name. Generally XAI refers to AI that can explain its reasoning process. Most AI systems today cannot, or are very bad at it. Among other things, this makes them hard to trust. (Duh.) And therefore much less useful.
Meanwhile, professors on reddit say they can tell a student used AI because of 1) tired phraseology and 2) factual errors.
Apparently the existing AI solutions just make shit up.
AI chatbox's penchant for making things up sounds rather like the chatty rationalistic left brain hemisphere after a stroke.
Unmoored from the more intuitive right, it'll propound all kinds of stupid shit, using sound logic and language, but won't even recognize half its own body.
Sort of a male/female dichotomy.
You have two AIs. One is an "AI that is explicitly fascist and white supremacist". The other is an AI that is explicitly queer, pozzed, government funded, supported, universally praised by "all the right people", and gets its electricity through a buttplug attached to a wall outlet.
At this point - knowing what you know - which would you trust more to give you a workable answer based on the truth?
There's a tendency for people to use the wrong word when they want to add emphasis: I can't seem to break my kids from the habit of saying "literally" when they mean figuratively. On this post's subject, I'm pretty certain that the WaPo commenter meant implicitly, not explicitly. If it was explicit, they would have said that they intend this AI to support a fascist and white supremacist agenda. Which is somewhat unbelievable, in that Musk, along with most Americans think "white supremacist" is a BS term:
Althouse describes The Post's comments section as a "pile." Glad she finally gets it.
"Radio preachers in the early 80s, one in particular, were warning about artificial intelligence, only this preacher called it "A One" because he'd never heard it pronounced."
Led to the Great Steak Sauce Freakout of 1986.
“They can’t get lane keep assist to work right. Full autonomous driving is looking like a scam…but you do you.”
Bought a 2023 Audi Q7 this spring. Took off on a drive from PHX to LAS (en route to MT). Had to pull over at maybe 50 miles for my partner to throw up. Luckily, it was slow enough coming on that we could find a place to pull off first. It was the Lane Assist. It was a two lane each way, divided highway, with almost no traffic, and curves. Speed limit was a nice conservative 75 or so. Turns out, that to drive it comfortably, at the speed limit, you tend to cut the corners of the lanes a bit. My problem was that Lane Assist, from the factory, kept you in your lane. Period. Or really tried to. Which meant jerking you back into your lane, when cornering too fast (which was what made my partner car sick), and increasingly resisting lane changes without signaling - even if you are the only car in 10 miles.
Pulled over on the side of the road, after she had hurled, after I dug into the problem, I was able to detune Lane Assist enough to comfortably proceed. Most of the time. It does give me a nasty message sometimes when drifting lanes, but I don’t mind some tactile reminders. I would prefer to turn it off, but that is apparently, not an option in that car (I was able to turn it off on a couple rental cars last year). But that could be a result of meddlesome federal regulations (such as the one that mandates that the rear view camera be disabled over maybe 5 mph - my Tahoe has it turned on all the time, and I love it).
I think that the thing that is hard with self driving cars is the transition between non-autonomous and fully-autonomous. Musk apparently runs fully autonomous Teslas running in his Boring Company test system, under Las Vegas, just fine. Computers deal just fine with other computers. And humans deal decently well with other humans. But humans aren’t as rigid in their programming as computers are. The authoritarians running our government and much of our industry are just fine with giving control of society over to machines, because those machines would be programmed to do their will. Intelligent people, who aren’t a part of that ruling elite, aren’t fine with it.
Temujin said...
Let WaPo and it's readers check off boxes for skin color and vaginas, because, you know...humankind's future depends on counting skin color and vaginas.
I'm pretty sure you're supposed to be calling them "bonus holes" now...
Temujin said...
"Let WaPo and it's readers check off boxes for skin color and vaginas, because, you know...humankind's future depends on counting skin color and vaginas."
I'm pretty sure you're supposed to be calling them "bonus holes" now...
fascist and white supremacist." Meaningless.
Musk's AI team is "all male"? In a world where a Supreme Court justice can't define female, I'm gonna need to see how the WaPo verified that fact. I highly doubt that the reporter asked each team member their pronouns (and that's not even getting into how one counts they/their identifiers).
"We hate Musk."
I've been hanging around YouTube lately. Given the predilections of the Google/YouTube crowd I've been surprised at the preponderance of videos of conservatives (Shapiro, Owens, Walsh, etc.) feasting, albeit politely, on snarky progressives.
Reading this post it occurs to me that "I hate" and "We hate" do not make for articulable points of view. I am never certain, but it also appears to me that most of the snarks identify as women, although some may just be girly men. Just sayin'.
Those commenters are really like AI; programmed to think and act a certain way with no self awareness.
"Lane assist" Christ, I just figured out what was wrong with the Hyundai car I rented for a month in Ireland this summer. Flippin' lane assist! I thought it was just an annoying beeping when I crossed the line (real or imaginary), but I also noticed that the wheel was getting jerked around a bit on winding roads (almost all of them...). If you've spent any time driving the narrow winding roads of Ireland or the UK, you can imagine how pointless this feature is. In AZ both my cars are much too old to bother me with this sort of nonsense.
“ Blogger wild chicken said...
Meanwhile, professors on reddit say they can tell a student used AI because of 1) tired phraseology and 2) factual errors.
Apparently the existing AI solutions just make shit up.”
Some of the AIs are “trained” by having them read everything they can find on the Internet, both fiction and nonfiction. Those AIs simply don’t know the difference. They’re just as likely to cite something from a novel or kook website as from a factual source.
Lane Assist in a rented Nissan had me on edge for a couple of days in '18, driving to Savannah GA and back. I'll have to ask about it if and when I rent a car in the future. It sucks.
"Math is hard."
By which SHE really meant, "Basic arithmetic is hard."
Bruce Hayden said...
It was the Lane Assist. It was a two lane each way, divided highway, with almost no traffic, and curves.
If you happen to come up on gilbar on an Iowa two lane road, you'll quickly notice that he puts on his turn signal for curves. Why? Because Then he can cross over the centerline without his pos Honda HR-V squawking at him. Y'all might want to try this yourselves
I drove a rented Audi with lane assist and had the problems mentioned above. But I also found that the car couldn't figure out what to aim for when lane markings were temporarily missing
as happens in rain when water pools on the road. And I've read that driverless cars are being completely halted in San Francisco by protestors who put traffic cones on the car hoods. Evidently the tiny car brain interprets this large, suddenly appearing cone image as signaling STOP STOP STOP. In short, these cars are easily fooled and that makes unassisted driving dangerous.
PS You guys may think you are making calm, rational statements about women. You come across as "mad as hell and not taking it any more." I don't blame you based on the propaganda pumped out, in the last five years especially, against you. But surely you don't really think that Rachel Levine, Sam Brinton, and Dylan Mulvaney and their ilk are representative women? Heat of the moment?
I drove a rented Audi with lane assist and had the problems mentioned above. But I also found that the car couldn't figure out what to aim for when lane markings were temporarily missing as happens in rain when water pools on the road. And I've read that driverless cars are being completely halted in San Francisco by protestors who put traffic cones on the car hoods. Evidently the tiny car brain interprets this large, suddenly appearing cone image as signaling STOP STOP STOP. In short, these cars are easily fooled, once you undertsand their guidance systems and that makes unassisted driving dangerous.
PS You guys may think you are making calm, rational statements about women. You come across as "mad as hell and not taking it any more." I don't blame you based on the propaganda pumped out, in the last five years especially, against you. But surely you don't really think that Rachel Levine, Sam Brinton, and Dylan Mulvaney and their ilk are representative women? Heat of the moment?
I wouldn't bet against Musk in his vision of what he wants xAI to be. He's taking it into the right direction; AI apps now are the equivalent of what Space Invaders was for video games: Nice start, but room for several worlds worth of improvements and growth.
"Just what we need, AI that is explicitly fascist and white supremacist."
Which is to say, an AI based in reality, instead of in left wing lies
Musk said the fundamental goal of the company is to understand 'what the hell is really going on.'
"Trans women" are men, Bidenflation is horrible, and the Covid shot is worthless for most people under rite age of 60.
Yes, with a reality like that, one understand why Leftists hate anyone who wants to find out the truth
Lane assist in our new Highlander causes the car to weave back and forth between the left and right lane lines, unless I consciously resist its prompts. It annoys me but not my wife. That's the way she drives anyway. This is to be her car. I still have my truck.
wildswan,
Those guys aren't women, they are guys. And insane (or hyper-trolling, maybe.)
And what folks are talking about here is really the aggregate; of course there are outliers. My take on it is this: You can't build a society, a culture, a civilization, on the exceptions.
"Musk's team of all-male math, AI, and engineering experts..." The horror!
Looks like a great opportunity for George Soros to put together a team of all female under 50 non-while male social justice sustainability and women's/ethnic intersectionality PhDs to go out and clean Musk's clock with a new AI they develop!
You go, Girlz and George!
Elon, OTOH, made the first profitable mass-market electric automobiles, figured out how to put gigantic internet communications satellite constellations in orbit with reusable rockets (and probably saved the Ukraine by giving them free access to that system), ferries Astronauts to the ISS when NASA can't, and just might get the first humans onto Mars.
Soros, OTOH, figured out how to hammer the British pound via currency trading (and Dem fundraisers consider that a much greater contribution to humanity).
So I ain't betting money on the success of the Soros group...
Lane assist in my Porsche Macan sucks.
Adaptive Cruise Control, OTOH, is remarkably good.
I'm thinking that if an autopilot (AI) can presently figure out how to fly a 450,000 pound 767 from Hawaii to the west coast without crashing, then pretty soon AI will be able figure out how to drive a car better than a stoned or drunk teenager on I-80.
"all-male"
Like there's something bad about males who like math and tech, or they should somehow feel bad that more women don't like what they like.
"Fascist" now only means "those other people in power."
Open AI seems to have a left bias, so there's another platform that musk is repairing.
Free speech is Fascism. got it
If the problem is that AI has trouble distinguishing between fact and fiction, I suspect that it is just a matter of time (probably a very short time) before we get "approved" or "curated" AI put out by various corporate and governmental outlets. Sort of like what we have now in the "news" and "information" sectors, except moreso.
I am going to go out on a limb and predict that those outlets which are curated to rigorously exclude "made up" "facts" and information will very quickly be labeled "right-wing" and "fascist".
“ Moondawggie said...
Lane assist in my Porsche Macan sucks.
Adaptive Cruise Control, OTOH, is remarkably good.
I'm thinking that if an autopilot (AI) can presently figure out how to fly a 450,000 pound 767 from Hawaii to the west coast without crashing, then pretty soon AI will be able figure out how to drive a car better than a stoned or drunk teenager on I-80.”
For my wife’s Honda CRV, I turn off lane assist because it’s annoying. When I pass a semi, I like to give a little extra distance because of the turbulence, but the lane assist tries to force me closer to the truck. I don’t like that.
I’m a private pilot, and in my opinion automating control of an airliner is a much easier challenge than making a self-driving car. Airliners fly steady speeds at assigned altitudes, with air traffic control and the TCAS system to assure separation. There’s nothing like that for cars. Throw in road hazards, idiots drunk, stones, or on their cell phones and the highways are a more complicated and chaotic environment than the airways.
"The billionaire and his team ... "
Forbes reported in April that there were 2,640 billionaires in the world, and we all know that Elon Musk is one of them so isn't there a more interesting descriptor that would apply here? Electric car innovator? Commercial Rocket founder? Twitter savior?
Post a Comment