It’s not easy to talk down a $500 billion initiative, but it’s happening. The Stargate Project is ambitious, and arguably a good strategic move. It’s an America First thing, to maintain the US lead in AI.
The trouble is that nobody quite believes it. The AI of just about everybody but China is being ignored.
Stargate is a broad-brush approach to the tidal flows of new tech AI that are generated on a daily basis. The credibility at the coal face comes from the direct involvement of OpenAI and other sector heavyweights. Microsoft is partnered with OpenAI.
Then there’s the money. A round number doesn’t mean much to hardheads in the tech sector. It’s ballpark at best and so far, short on specifics. Musk claims they don’t have the money, but it’s not too hard to see cash moving into a project with US government backing and support. Nvidia is also in the mix as a supplier.
These companies have the market clout to generate cash if needed. That’s particularly the case when so much investment money has poured into the sector in the last two years.
The general bitchiness is pretty obvious. Musk doesn’t like it for various reasons, stated or otherwise. Trump is doing his usual hype, which is always monotonously similar whatever the subject. “So what?”, you may well ask.
AHEM.
Never mind the gossip. OpenAI has been pretty open, excuse the expression, about its progress to AGI, the next level AI. They’ve set the benchmarks high, despite the market and the general babble from the sector. That’s a necessity, not false modesty.
A $500 billion figure doesn’t make it any easier. Sector expectations are always overstated, but delivery on investment is much less ambiguous. Failure to deliver is not an option.
China’s AI is another factor. There’s a deeply ironic tale here in that the much-reviled Huawei is providing the chips. Nobody seems too sure exactly how close China is to US AI tech levels. The assumption is that it’s at least a bit behind.
All of which blather leaves out the rest of the world. Most technologies aren’t purely national. France, the UK and EU are all making various noises about AI at different levels.
Then there’s getting AI into the physical world as a working thing. Codes of practice are likely to be sticking points. The legal issues for AI could be encyclopedic on any noun and verb.
What’s likely to happen?
To start with, forget the $500 billion figure as a yardstick. The IP alone for AGI will be worth far more, if and when it works and delivers value.
The economic effects of AI are mainly fortune-telling at this point. Not enough thinking is going into how to “install” AI in the real economy. This total lack of attention to the obvious is likely to turn into a hyper-expensive Wile E Coyote cartoon.
The “fire everybody” motif is also totally unrealistic. Wherever money and hard assets move, oversight is required. You have to talk to somebody. You have to negotiate with people. You need someone minding the store and making sure things work.
The skill sets at all levels of business are also hardly up to scratch. How many managers, accountants, etc, can believably say they know how to handle whatever comes out of the machines? Even basic prompts for AI can get messy and costly.
Speaking of cost – The dogmatic emphasis on the costs of AI processing times can’t work. If you’re going to have a fit over the cost of every process, it’s counterproductive. Maybe a new approach to energy efficiency will help, but it’s still at the startup stage.
(Excuse me a bit of sarcasm at the world suddenly “discovering” the need for energy efficiency all these years later. Basic engineering, bozos.)
Stargate could be a great move or a huge comeuppance. We’ll see.
___________________________________________________
Disclaimer
The opinions expressed in this Op-Ed are those of the author. They do not purport to reflect the opinions or views of the Digital Journal or its members.
