top of page

DeepSeeks Wild Ride

  • Writer: Ray Alner
    Ray Alner
  • Feb 4
  • 2 min read

Market Shaker

Well, well, well. It seems we have a competitor.

That didn’t take long.

DeepSeek, a company based in China, claims it made a model 20-40x cheaper than ChatGPT and the short sellers wipe billions off the market.

Pretty sure everyones heard of the story, so I’m not going to belabor the point. Lets get into what I think about this ridiculousness.

The Overreaction

I’m a bit confused by the response the financial market had when that article was released.

China just had to say the words “We made a better genAI for 40x lower cost”, not even prove it and the markets decide that is enough for a massive market correction.

Is the market going to be rattled every time someone OTHER than our best tech darling, OpenAI makes a better model, or is it just because a company in China made the model?

Have we consolidated and inflated values so much of our financial wellbeing on one single field and one single company that leaps IN THE SAME FIELD from numbers that haven’t even been independently verified will shake us this badly?

What if some other innovative tech from another company comes out and proves a big success? Are we going to jump ship even quicker this time? Are we fine with this potentially causing a recession because we have all our eggs in one basket?

The Innovation

Pat Gelsinger said what many of us are thinking, that when we limit chip access through trade restrictions we don’t decrease their capabilities to produce a better product, we just allowed them to come up with a creative solution to a difficult problem.

In the long run, no one is going to lose, even if it is another model. This alleged reduction in cost only means that someone CAN get the cost down 20-40x, giving US companies an even better chance at reducing the cost of generative AI, which if I’m not mistaken, is estimated to lose OpenAI 5 billion this year.

There's much more to say about this, like the fact the model may have been trained on ChatGPT or similar datasets ChatGPT was trained on, or they had a data breach, or senators trying to restrict the import of models. These discussions don’t solve the problem that innovation can happen from anywhere, and anyone, and just because DeepSeek may have flamed out quickly, doesn’t mean we should react like the sky is falling every. time. someone. ELSE. innovates.

Comments


bottom of page