For all the hype, babble, and actual achievements, America’s new cash cow AI has just had a jolt of very expensive reality. Despite the Stargate Project, designed to maintain America’s edge in AI, China’s DeepSeek has just punctured the balloon.
Somehow, no doubt in a flurry of macho meetings for tiresome rodents, the idea of American exclusivity in AI was taken seriously. No form of tech is exclusive and never has been. Fire and the wheel are good examples.
There are hordes of AI businesses trying to create a space for themselves in this market. The highly publicized variety of AI is nothing like the whole story. The next generation of AI hardware is also up for grabs.
Then there’s theft. China is the bogeyman, but does anyone have the inbuilt level of extraordinary stupidity required to believe that nobody else is swiping AI tech?
Tech media has been saying for a long time that the hype has been out of control in the investment markets. Billions of dollars have found obliging rabbit holes in this techno-mythology.
DeepSeek also had the market perceptiveness to bring some credentials with it. DeepSeek is an actual open source AI. It’s also affordable and pretty cheap. Current gossip is that it’s better for technical tasks, while ChatGPT is more chatty… hard to avoid that analogy because it’s a better chatbot.
The markets, meanwhile, had a demure fit. Nvidia, long said to be due for a major correction, corrected itself and accounted for most of the $1 trillion downward move. Other tech stocks were also hit, and tech indices must be feeling pretty queasy, too.
This undignified stampede caused by the mere presence of a credible competitor has a much deadlier if almost too obvious message.
This is a true global market.
America is not the only player.
Big money wants returns.
Investments in AI must deliver hard dollar values.
Another good idea is not bleating about the cost of AI processes and doing something useful, like making them market-ready and cheaper.
One of the reasons for the instant and ongoing extreme skepticism in the tech sector is that this drivel been hyping the wrong things. The chat-related functions are all just large language models hooked up to old software ideas and chatbots. The market has been seeing what it wants to see, as usual.
You can write an essay based on a search, sure. So what? This is almost literally playing with alphabet blocks in kindergarten at a cost of billions of dollars.
The next generation of AI will make a big leap in capabilities. From what I’ve seen, AI analytics are generally pretty good. I have also caught a few indicators of bias. I’d have to see a lot more and cross-check to spend a cent on any of it, but OK-ish.
For example – AI peer review is an interesting option. How does an AI oversight another AI? Does that add a level of difficulty that an AI will respect? If you tell an AI its work will be checked, does it behave differently?
I mention this because the market doesn’t seem to know what constitutes good performance by AI. That’s the incredible weakness DeepSeek has just exposed.
You’re not buying an iPhone, an air fryer, or some showy peripheral.
You’re buying a lot of money’s worth of something, and you don’t know what? Fix that.
___________________________________________________________
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.
