I think Musk wants Twitter to use for AI training

[intrepidity]

Last edited Sun Nov 6, 2022, 03:05 PM - Edit history (1)

He is deeply into AI, machine learning, neural nets. Those things *all* need training sets from actual humans.

I do believe that his purchase has more to do with harvesting data than anything else. We'll see.

ETA on 11/4: Just saw a tweet from Elon saying

"Because it consists of billions of bidirectional interactions per day, Twitter can be thought of as a collective, cybernetic super-intelligence"

and if that isn't the biggest clue that I'm correct, I don't know what else to say.

[mahina]

1. Concur

[canetoad]

2. Rise

Of the Musk Bots.

[GoneOffShore]

3. Or the MuskRats.

[cstanleytech]

4. Dont you mean MuskoRats?

[Bernardo de La Paz]

5. Twitter and all Big Data uses AI and are all deep into AI. Further, he does not need Twitter for AI.

He can do R & D on AI more easily on research platforms. He already is.

He does not need the headaches of running yet another high profile company just to run AI.

He bought Twitter because he ran his mouth about censorship and backed into it. Then he tried to back out and couldn't break the commitment. So he was forced to follow through or he would have been liable for stock price losses. He already had problems with the SEC for running his mouth and affecting his stocks. He could not risk more SEC trouble or he might have been forced out as CEO of his companies.

He ran his mouth because of his ego and scored an own-goal against himself.

[intrepidity]

6. You'll know it when he starts talking about the AI content moderation

He'll describe how they do a superior job etc, but he'll need deep access to all of Twitter in order to sufficiently train it--not just skimming it from what gets posted.

Twitter will be the proving grounds. Whatever he learns there will underpin a next (or current) project.

Why *wouldn't* he do that?

[Bernardo de La Paz]

7. Of course he will say that. But it is NOT why he bought Twitter. . . . nt

[intrepidity]

8. I see it as a very calculated move.

Almost everyone thinks as you describe, which makes him seem like more of a buffoon and less of a threat. That amount of data and power in one man's hands *should* look more like a threat than a joke. I don't think he has any sinister intent, at least not yet. Time will tell.

[Bernardo de La Paz]

9. Was opposite of calculated. That's why he fought hard in court to back out

Now that he is forced to go in, of course he will make the most of it that he can.

But it was NOT calculated. Events and facts show otherwise.

[JohnSJ]

10. Well, he doesn't have my data, and he sure as hell isn't going to get it

[JI7]

11. I think it has to do with China

[crickets]

12. China, Russia, USA, AI, and huge ego. nt

[PortTack]

13. He wants to spread disinformation for the right..my opinion

[yardwork]

14. Harvesting data. I just posted that in another thread.

That's where the money lies.

[intrepidity]

16. But not just harvesting it for profit, per se

but to use for further development of his AI projects, which is where much of his interest is focused. Transhumanism, colonizing Mars, etc are projects that will benefit greatly from advanced machine learning. And I believe he sees in Twitter a vast resource of training data. Figuring out how to eliminate bots and spam will not only improve the service, but will further refine the training data. It's what I would do if I were him.

[Algernon Moncrieff]

15. That actually would make sense

A lot of what Tesla does, I'm told, is more about data gathering than building cars

[intrepidity]

17. I could see that

Especially with the self-driving vehicles. I mean, that is machine learning in action.

[intrepidity]

18. Updated OP to include Musk tweet

Link to tweet

[tinrobot]

20. Well... It's only "intelligent" when people use it.

At this point, he's trying very hard to make the opposite happen.

[tinrobot]

19. I think 90% of it is pure ego, but he did move some Tesla AI people over to Twitter

Mostly people on the Autopilot team, who are deep into AI. So, you may be on to something:

https://www.cnbc.com/2022/10/31/elon-musk-has-pulled-more-than-50-tesla-engineers-into-twitter.html

[Emrys]

21. I believe that's one of the useful applications data miners have in mind.

They need to be careful about the old garbage in, garbage out problem, though, which may mean non-rightwingers and non-absolute nutters won't be deliberately driven off the site wholesale:

Microsoft’s AI chatbot turns sexist, racist

In 2016, Microsoft launched an AI chatbot called Tay. Tay engaged with Twitter users through “casual and playful conversation.” However, in less than 24 hours, Twitter users manipulated the bot to make deeply sexist and racist remarks.

Tay leveraged AI to learn from its conversations with Twitter users. The more conversations it had, the “smarter” it became. Soon, the bot began repeating users’ inflammatory statements, including “Hitler was right,” “feminism is cancer,” and “9/11 was an inside job.”

As the debacle unfolded, Microsoft had to pull the plug on the bot within a day after its launch. Later, Peter Lee, Microsoft’s Vice President of research, issued an apology, stating, “We are deeply sorry for the unintended offensive and hurtful tweets from Tay, which do not represent who we are or what we stand for, nor how we designed Tay.”

https://www.jumpstartmag.com/ai-gone-wrong-5-biggest-ai-failures-of-all-time/

I'm not clear Musk is familiar with that experiment, but it may be drawn to his attention as events unfold.