BwazeIs it just me or is it really weird to write a short article about Musk venturing into AI, but not even mentioning Elon Musk's call to put a stop on AI development for at least half a year - less than two weeks ago?

I already mentioned it in the previous story about the data center GPU order - it's nothing really new at this point, apart from some of the big newspapers reporting on false signatures.

Only a few weeks ago Musk co-signed a letter in which he and more than 1,800 others demanded a six-month pause in AI research. It later emerged that some of the signatories were fake.

It's been long established that the fella is all about big bungles and contradictions. I thought that the Neuralink connection was more interesting, especially with his corporate heavy being involved in the formation of this AI startup.

Why_MeHow did Musk weaponize social media when he was the one who exposed US govt. - big tech collusion.

The same Musk who asked Taibbi to release the Twitter files because he wants independent journalism and then blocked Taibbi from doing independent journalism by blocking his (and every other independent journalist’s) substack? Free speech and independent media on my platform, but nowhere else!

Musk didn’t even read the Twitter files, by his own omission, then went on to tweet about government interference in the laptop story while Taibbi reported and tweeted that he found no such evidence.

Miranda Devine, who broke the laptop story, said there was no smoking gun. Every legal scholar said this was not a free speech issue, but a content moderation one (Twitter is a private company, after all). Wired called it “transparency theater,” but the intelligencer put it best,

“[the reporting is] saturated in hyperbole, marred by omissions of context, and discredited by instances of outright mendacity…best understood as an egregious example of the very phenomenon it purports to condemn — that of social-media managers leveraging their platforms for partisan ends."

Hilariously, Dorsey asked Musk to make all of the Twitter files public in the interest of transparency. Yet, they’re not. Free speech as long as it fits my narrative!

The medium is indeed the message, and in this case it’s the cesspool of social media and a billionaire moron with clear partisan objectives. Political theater at best, bad faith argument at its worst.

Can’t wait until the AI moderators are released and this idiot’s reactions, should be fun. Hopefully we get screenshots of the chats.

Why_MeHow did Musk weaponize social media when he was the one who exposed US govt. - big tech collusion.

Hi,
Indeed twitter was silly stupid before anyway.
Musk is an improvement no matter the bad he's done nothing is as bad as before even facebook nonsense is as bad as twitter was.

Vayra86Attention. Its narcism, everything else is secondary. Fits right in with Bezos, Trump, Putin, etc.


Lets agree to disagree :) That 3% of the worlds population in top wealth isnt government though...

Think the new term is Influencer :laugh:

claesThe same Musk who asked Taibbi to release the Twitter files because he wants independent journalism and then blocked Taibbi from doing independent journalism by blocking his (and every other independent journalist’s) substack? Free speech and independent media on my platform, but nowhere else!

Musk didn’t even read the Twitter files, by his own omission, then went on to tweet about government interference in the laptop story while Taibbi reported and tweeted that he found no such evidence.

Miranda Devine, who broke the laptop story, said there was no smoking gun. Every legal scholar said this was not a free speech issue, but a content moderation one (Twitter is a private company, after all). Wired called it “transparency theater,” but the intelligencer put it best,


Hilariously, Dorsey asked Musk to make all of the Twitter files public in the interest of transparency. Yet, they’re not. Free speech as long as it fits my narrative!

The medium is indeed the message, and in this case it’s the cesspool of social media and a billionaire moron with clear partisan objectives. Political theater at best, bad faith argument at its worst.

Can’t wait until the AI moderators are released and this idiot’s reactions, should be fun. Hopefully we get screenshots of the chats.

The NY Post broke the laptop story and they were banned from Twitter for posting it. Remember when 67 so called ex Intel officials said the laptop was Russian disinfo only to have the NY Times tell us the laptop was legit two years after the fact. Yes we remember. As far as Substack goes, Musk claims they were trying to steal Twitter's code. No idea if that's what happened but that's what Elon claims. Dorsey was proved to be a liar on the Joe Rogan show. If not for Musk purchasing Twitter the US govt. would still be going full Soviet on us.

Why_MeThe NY Post broke the laptop story and they were banned from Twitter for posting it. Remember when 67 so called ex Intel officials said the laptop was Russian disinfo only to have the NY Times tell us the laptop was legit two years after the fact. Yes we remember. As far as Substack goes, Musk claims they were trying to steal Twitter's code. No idea if that's what happened but that's what Elon claims. Dorsey was proved to be a liar on the Joe Rogan show. If not for Musk purchasing Twitter the US govt. would still be going full Soviet on us.

See, too much 'online news' gets you this deep into the rabbit hole. Dude, what?! None of what you're saying makes even a remote kind of sense.

Did you watch Chernobyl, the series? There's some golden dialogue in it; 'if all you hear is lies, how can you still know the truth'

None of this 'news' is truly relevant other than for serving another man's agenda. And it ain't the government agenda. Its the one of a dude with a big mouth who wasted over 20 billion on a social media service. Poof, thin air. If that man 'claims' anything to win souls, you should know what it means. It serves to secure the remaining 20 odd billion.

Just step out of this cesspool. Its all poison, relevance zero and every social media user is the product. Your clicks provide credibility to absolute gutter trash nonsense to gain other people money & influence.

Joe Rogan et al. are more of the same btw. This is the new tabloid press. A hype a day keeps the paychecks coming in.

Vayra86See, too much 'online news' gets you this deep into the rabbit hole. Dude, what?! None of what you're saying makes even a remote kind of sense.

Did you watch Chernobyl, the series? There's some golden dialogue in it; 'if all you hear is lies, how can you still know the truth'

None of this 'news' is truly relevant other than for serving another man's agenda. And it ain't the government agenda. Its the one of a dude with a big mouth who wasted over 20 billion on a social media service. Poof, thin air. If that man 'claims' anything to win souls, you should know what it means. It serves to secure the remaining 20 odd billion.

Just step out of this cesspool. Its all poison, relevance zero and every social media user is the product. Your clicks provide credibility to absolute gutter trash nonsense to gain other people money & influence.

Joe Rogan et al. are more of the same btw. This is the new tabloid press. A hype a day keeps the paychecks coming in.

Point out one part of my post that you quoted that isn't the truth.

So, total side note but... I've started playing Stellaris. And I can't help but notice that Elon Musk's logic is just bullshitting his way through bad Stellaris fanfiction.

Stellaris is a fine game, but it has a number of "bullshit" issues because... its a video game. Like, AoE2 has swords made out of gold, or Factorio allows you to build a circuit board out of iron. That sorta thing. Things that make sense "in game universe" but you kind of ignore them because we all recognize that video games have limitations, and we're just playing the game for fun and not for anything serious.

This AI issue? Just the AI Uprising event from Stellaris (stellaris.paradoxwikis.com/Situations#AI-Related_Incidents).

Trying to terraform Mars because "Earth is Dying??" (but not having the option to Terraform Earth?? to fix our own world?) That's just how the video game works, you can only terraform uncolonized planets in Stellaris, its not an option to terraform an already occupied planet.

"Energy Backed Currency" ?? That's just Stellaris energy credits.

Etc. etc. By accidentally finally playing Stellaris, I think I finally understand Musk. He's just a Stellaris fanboy it seems. Including all the bullshit and stupidity inside of the game, but now applied to the real world (where things won't work the way he thinks).

-----------------

That being said, I'm not sure if Elon Musk even needs to solve the AI problem. He's just a Stellaris fanboy, always has been, and tries to attract others into enjoying Sci-fi concepts in the real world. That's his brand. He's not actually about solving problems as much as he is about just doing SciFi things (whether or not they make sense or work is irrelevant). There seem to be enough fanboys who will give him money (ie: buy stocks or bonds, or his products) over this alone.

From that perspective, spending $20 Million on a bunch of "AI Sounding Technology" is all Musk needs to raise more money. Its like Solar Roofs, or Tesla Bot / Optimus, or "Full Self Driving". None of these techs are going to happen, but its just part of the grift to raise more money from sci-fi fans. IE: I'm sure Elon Musk will raise $100 Million from fanboys, recuperating the $20 Million GPU costs.

Why_MeYes because the government colluding with big tech in order to suppress free speech isn't big deal .... said no American ever.

Not going there other than you just met one.

Why_MePoint out one part of my post that you quoted that isn't the truth.

Pretty much all of it. You honestly aren't even intelligible other than a loud "Hunter Bidens Laptop!" screeching sound.

Beside all that, this is incredibly off topic. Just. Stop.

On a side note, I'm making a 2nd post on a different issue completely separate from Musk and focus on the technical problem here.

LLMs are fancy autocomplete. The sooner people realize, the sooner we all will be better people. Look at the "shape" of the artificial neural net and think about how it works. I'm not saying this is how it will always work, but... its how LLMs (like ChatGPT4) currently works.

In practice, this means that LLMs are exceptionally good at choosing "the next word to say". That is 100% where they shine, and its incredible when you use LLMs for their designed purpose. Writing some code? Well, here's fancy autocomplete that will generate 50 more lines for you, its probably not what you want, but maybe it gives you an idea of where to go. Awesome, Github Copilot LLM doing great.

Writing fiction? Well, have GPT4 finish your next paragraph. Awesome. That's how fancy autocomplete works, it generates a paragraph based on the words you already wrote... except this thing is good enough to analyze the last 10,000 tokens (~maybe the last 5 or 6 pages of text) and use all of that to calculate the next word. Incredible for sure.

-------------

Where do people start going wrong with LLMs? Research, sentiment analysis, etc. etc. Or other such classification tasks. Its not like this AI knows right from wrong, truth from lies, or anything. Its just writing the next word based on the words that have come before. It sounds like Reddit because its been trained on Reddit (iE: where to start a new post, where to stop a post, when to respond, how to respond). It sounds like StackOverflow because its been trained on StackOverflow.

But that doesn't mean its actually conducting research. It means that its autocompleting the conversation. Its an incredible way to write a "Chatbot" (and the innovation here is that "ending your current thought" is part of the discussion. So ChatGPT knows when to stop talking and wait for a response... because its been trained to do so).

Thanks @dragontamer5788 for bringing it back on topic, i.e., AI. Kudos, dude.

AI is the topic. Try it someday.

Does anybody have a link to that Google paper that is supposed to be the base for ChatGPT? I read an article recently about it but I can't remember where.

unwind-protectDoes anybody have a link to that Google paper that is supposed to be the base for ChatGPT? I read an article recently about it but I can't remember where.

I don't know what paper you're referring to, but ChatGPT is using the transformer model intruduced by google in 2017

en.wikipedia.org/wiki/Transformer_(machine_learning_model)

trsttteI don't know what paper you're referring to, but ChatGPT is using the transformer model intruduced by google in 2017

en.wikipedia.org/wiki/Transformer_(machine_learning_model)

Yes, this is what I mean. I guess the paper is Ashish Vaswani et. al. 2017 "Attention Is All You Need."[

It even looks pretty readable. arxiv.org/pdf/1706.03762.pdf

What's with the politics in this thread?

unwind-protectWhat's with the politics in this thread?

Apologies. I don't know how they got here and was merely responding. I actually appreciate the LQs because none of it should have ever been here at all, including my own, but that would require the subject to not have been raised in the first place.

AI at this point in time isn't even a useful tool in my eyes. It messes up with extreme confidence (and defends its messups with an almost irritated tone when questioned) far too often to be useful. But because it IS a buzzword now it is only a matter of time until it's used for something serious, and hurts people, if it hasn't already. (I'm looking at you, "early cancer warning" AI imaging).

The only real fun I've had with it is creating AI "art" of bizzare word prompts.

R-T-BAI at this point in time isn't even a useful tool in my eyes. It messes up with extreme confidence (and defends its messups with an almost irritated tone when questioned) far too often to be useful. But because it IS a buzzword now it is only a matter of time until it's used for something serious, and hurts people, if it hasn't already. (I'm looking at you, "early cancer warning" AI imaging).

AI, in its current form, is highly effective at raising money. All you gotta do is say that your company is an AI company, and then the Venture Capitalists shove a ton of money in your direction.

With all the money being thrown in AI's direction, someone will inevitably do something and it will be considered a "success". Well.... maybe not necessarily. Metaverse had what? $40+ Billion thrown at it? At least, the "chances of success" are probably higher.

dragontamer5788With all the money being thrown in AI's direction, someone will inevitably do something and it will be considered a "success". Well.... maybe not necessarily. Metaverse had what? $40+ Billion thrown at it? At least, the "chances of success" are probably higher.

I dunno. What defines "success?" What you just described was basically the word "blockchain" 2 years ago...

R-T-BI dunno. What defines "success?" What you just described was basically the word "blockchain" 2 years ago...

On the scale of "hype projects that got a ton of money", we're looking at XML, OOP, UML, Agile Programming, Test Driven Development, Cloud, "As a Service", IOT, Blockchain, Metaverse, and now AI. Well, ignoring the "Expert Systems" AI Boom of the 80s of course, which also was a ton of money into random things.

Bonus points if someone (IBM) tries to use 3 or 4 words at the same time (IOT Sensors giving data to your private Blockchain saving the data to the cloud accessed through our platform-as-a-service). Its so stupid its not even satire: www.ibm.com/cloud/architecture/architectures/iot-blockchain-cold-chain/

Anyway, "what is a success"?? Anything about... say... XML's level or better in my book. Used in .docx and largely forgotten about otherwise (as JSON turned out to be more convenient in practice).

You probably forgot about "social media" ~ arguably the second biggest curse(?) of 21st century after these of course!

Gotta love them buzzwords :D

R0H1TYou probably forgot about "social media" ~ arguably the second biggest curse(?) of 21st century after these of course!

Ahhh... crap.

The "X App" is just an AI-powered blockchain payment processor social network programmed in an OOP language with a cloud-first REST API, isn't it?

Welcome to Silicon Valley folks. Its powered by buzzwords.

EDIT: "As a service", now that $8/month blue checkmarks are a thing. Are we missing any buzzwords here? Is there any room for anything else?

So when he called for 6 month pause in AI development he was just asking everybody to stop and wait for his AI company ?

r9So when he called for 6 month pause in AI development he was just asking everybody to stop and wait for his AI company ?

I don't think he was amongst those calling for a halt, was he?

It would surprise me if he could think that hard about consequences, frankly.

R-T-BI don't think he was amongst those calling for a halt, was he?

It would surprise me if he could think that hard about consequences, frankly.

He was one of the headliner's, I believe the theory originally was that he was salty - which by his public statements it's also true - about OpenAI turning into a for profit coorporation (technically there's a non profit and a for profit side of the business, the usual legal workarounds) and slowly being bought out by Microsoft - he was an investor in the beginning and invested in what was a non profit.

But yeah, we now see it was also pure hypocrisy and a long shot hope for time to catch up.

The dangers of AI is an interesting one, the advances are pretty incredible in all areas, reading about the award winning photo that was pure AI generated: www.bbc.co.uk/news/entertainment-arts-65296763 reinforces how quickly people can be fooled.

And with Russia proving to be pathetically weak bad guys, AI should be a something worth focusing on.