Elon Musk plans artificial intelligence startup to rival OpenAI | Hacker News Reader

Elon Musk plans artificial intelligence startup to rival OpenAI

ft.com

11 pointsmepian11 comments

I'm continually surprised the boards and/or the shareholders of his public companies have not sought to rein Elon in.

Even if we accept the commonly-quoted meme that his IQ is in the 150+ range (which I strongly doubt), a person simply can't be an effective CEO/president/whatever of 5+ companies.

I'm no longer a TSLA shareholder but if I were I'd be more than a little peeved that the CEO is so chronically distracted.

Well after seeing how Twitter is going, and the rumors about how SpaceX and Tesla have learned to “manage” him, it almost seems like the less time Musk spends focused on a company, the more freedom they have to operate successfully. Perhaps it’s a benefit, as the more companies he’s involved in, the less damage he can inflict on any given one.

Reminds me of the stories of Pixar and Steve Jobs. They basically had jobs-specific daily previews, and then a second set of previews so people could get stuff done.

That said, they both had vision.

There's no truer indictment of an executive other than their absence adding more value than their presence.

A company like google or microsoft can easily be thought off as conglomeration of 5 or more companies. Mostly in such cases job of ceo is resource allocation and strategic thinking. If those companies can have a ceo then elon can also try

Perhaps, he can fix auto wipers first ?

Auto wipers are technology of yesterday, technology of legacy OEMs, which are going bankrupt anyway. In Tesla we are working on laser wipers [0] which will ablate all particles and vaporize all the fluid from your windshield. It also might damage your retina but who needs a retina, when FSD is doing all the driving.

/s obviously

https://electrek.co/2021/09/08/tesla-patent-laser-windshield...

I cannot blame him. Elon Musk was a founder of Open AI and resigned in 2018. Open AI ironically titled given it's no longer open. He should have more than enough resources to make it work. Although Tesla is in many respects already an AI company. Why not just make an AI subsidiary from Tesla?

Explain what you mean by Tesla already being an AI company?

They've been training and running neural nets for awhile.

But even more significantly, they have the Tesla FSD chip series. Custom chips are immensely expensive to design, and even using that as a starting point for a training focused chip would put them ahead of most AI startups.

Don't get me wrong, I am no Tesla acolyte... but its a good fit.

They’ve been heavily involved in self-driving car research for a while now. While that hasn’t achieved everything that’s been promised as of yet, they definitely have a lot of ML expertise in house.

Reply on news.ycombinator.com