What if
It was not Tesla, but Twitter/X that built Dojo.
Musk missed a trick there, to transfer Dojo to Twitter/X, which then could have monetised it. What would the value of Twitter/X be then?
Tesla's forthcoming Dojo supercomputer could add as much as $500 billion to the automaker's valuation – which currently sits at $875 billion - according to a Morgan Stanley Research note. The financial services giant is bullish on Dojo as it feels Tesla's development of custom silicon for the super will have applications well …
Er, no. Musk might manage a stable of companies, but that doesn't mean it's legal to give companies in that circle preferential treatment. There's fiduciary responsibility to get the best deal if it's being offered for license or rental. If Musk owned all the companies outright, yeah, he could do that, but he has to answer to investors and shareholders.
On this iOS Google keyboard, Itanium or as the Register called it "Itanic" isn't even auto filled.
You should check Wikipedia which has a section about big name projections about the future and market predictions about it. We were supposed to run it on our desktops retiring x86
Nevertheless both Intel and HP were seriously managed companies by professionals. Not someone in a highly publicised mid-life crisis.
by the insiders, sorry scumbags on Wall St
A computer is just consuming energy unless it is doing something useful. So far that step is lacking in specificity. Naturally Lord Almighty Elon the Magnificent will claim all the praise for the work of others.
SOP really for narcissists.
So the company which year after year proclaimed that full self-driving capabilities for their cars was just around the corner, when the only thing Tesla FSD does reliably is hitting emergency vehicles, and which has a track record of over-promising in pretty much everything else (Tesla robot, Cybertruck) is supposed to gain >50% in value because it buys Nvidia GPUs to build a "supercomputer"?
But then, this is Morgan Stanley, who has been on the wrong side of almost everything it touches, so if they think that building this computer makes Tesla more valuable then it's probably safe to assume Tesla's stock is going tank even more than it already did in recent times.
But then, this is Morgan Stanley, who has been on the wrong side of almost everything it touches, so if they think that building this computer makes Tesla more valuable then it's probably safe to assume Tesla's stock is going tank even more than it already did in recent times.
The thundering herd are just doing what they do best. So a pump & dump operation that stays just on the right side of the law. But-
Twentyfive of these chips are meshed in a 5x5 grid using TSMC's chip-on-wafer-on-substrate (CoWoS) packaging tech to form the Dojo training tile.
I'm developing a 7x7 grid, except 7 of the chips in the tile are extra custom, therefore mine should be worth at least $1tn! Buyins are available to early investors who want to get on board the future, once I've figured out the safest jurisdiction to stash the cash or crypto. But I digress. Curious if they'll ever get this to work, but can't quite see the revenue streams given it seems to be a rather custom build. Then again, there are all sorts of neat photogrammetry applications that hold much promise for future versions of GTA. And of course it'll probably run Crysis just fine.
> And of course it'll probably run Crysis just fine.
Ironically, it probably won't - Crysis wasn't written to take advantage of multiple cores, so performance doesn't scale well at all...
https://www.dsogaming.com/pc-performance-analyses/crysis-remastered-suffers-from-single-threaded-cpu-issues-just-like-the-original-game/
Ironically, it probably won't - Crysis wasn't written to take advantage of multiple cores, so performance doesn't scale well at all...
Hmm.. Aha! 101 uses for a pseudo-AI! Speedrunning records! Multiple cores running multiple instances with some machine learning to test and optimise outcomes. Collect speed running record! But I'm not entirely convinced that justifies a $500bn valuation.
Yeah, surely if 'Dojo' cost ~£1Bn to develop, it adds maybe a few $Bn in value over it's lifetime, tops. Anyway, until Tesla start paying shareholders dividends, it's just a massive Ponzi scheme, and whoever is left holding stock when the music stops loses. 'Dojo; is just another monkey to crank the handle.
Dojo could see Tesla become an AI-as-a-service provider to automakers that need FSD capabilities
I don't understand : are the self-driving vehicles supposed to be always connected to this Dojo, and off-load the real-time calculations of its surroundings to it ? Or is this only for training, and the cars embed an actual local processor ?