You live by the sword, you die by the sword.
If you are in the business of producing cutting edge technology - you had best be constantly ready for someone trying to come along and eat your lunch. That is just the reality of the arena in which you operate.
This being the case, it maybe it shouldn’t come as such a massive surprise to us that it turns out the Americans are not the only folks who can innovate.
Deepseek, a Chinese developed large language Artificial Intelligence model bearing more than a passing resemblance to Chat-GPT, came to market on the 20th January and in the process kicked off a sell-off in some of the biggest US tech stocks last week.
It is my (fairly limited) understanding that it isn’t so much the capability of Deepseek that has caused panic on the other side of the pond - but the fact that it was developed at such a relatively low cost, estimated to be $6mn versus the $100 million it took to train Chat-GPT.
Nvidia had become the biggest company in the world by market cap because in the great global AI arms race, Nvidia were the only gun dealer in town.
Building generative AI models requires a lot of computer power. A lot of computer power requires a lot of computer chips. And the best chips are all sold by Nvidia.
But if Deepseek has shown that it is possible to build models of comparable quality, without the same kind of reliance on the best, most sophisticated hardware - then maybe Nvidia’s market position maybe isn’t as dominant as we thought. Which isn’t great news for the share price.
What is going to happen from here remains to be seen. But I will admit to being somewhat amused by the US tech bros, who espouse so vocally for free market innovation, throwing their hands up in the air and crying foul play whenever some competition comes onto the scene. Time to up your game lads.
This whole episode reminds us of a couple of things.
First, this is why we invest at all. The human race is amazing at innovating, either improving things or producing the same quality of good or service but at a markedly reduced price (as is the case here).
Although our living costs, the price of the necessities, broadly rise over time - technological development is inherently deflationary. We can get access to ever better technology at lower and lower cost.
The amount of computer power in your hand, right now, would have needed a storage unit thirty years ago. Isn’t that amazing?
My second thought is that this development provides further evidence that the best way to hitch a ride on this innovation, is to invest in the global equity market as a whole.
Because, innovation being the ever present force that it is, means that the leading companies will likely naturally always be shifting and changing.
It might seem difficult to believe given how dominant the big US names have become, but market history is marked by cycles where the leaders change and evolve. We very likely haven’t even heard of any of the “Magnificent Seven” of the future yet.
The great thing about buying the whole market, is that we can guarantee ahead of time that we will be in these future winners when they emerge.
On the other hand the downside of trying to be clever and trying to pick the individual winners is that, sometimes, life comes at you fast. Two weeks ago, Nvidia had a unassailable market position. Now it might not.
Share price equals earnings multiplied by narrative. For quickly growing companies like Nvidia, with huge enthusiasm baked into its share price - narrative is everything.
If that narrative moves against you, the shares are left floating in space like Wile E Coyote when he runs off the side of the cliff.
The final thought I had, and this is kind of my main point here, is that when these kinds of things happen - big news stories and world events, it is brilliant for bringing out the bullshitters.
You know the kind. Have to have an opinion on everything, no matter how ill informed. Generative AI, Middle Eastern politics, 1970’s discus throwing - no subject is beyond the bounds of their (mis)understanding.
Getting older is recognising just how little I know about the world.
An ever shrinking circle of competence is a good thing, because it keeps you on-piste. The bad stuff only really happens when you get out over your skis into a world that you don’t really understand.
I know almost nothing really about Artificial Intelligence, in the same way that I cannot possibly know today what impact a Trump presidency will have on the stock market and that is ok.
Good investing recognises our limitations and continues to paint within the lines.
Chat-GPT really is amazing, I pretty much use it interchangeably with Google now. But I don’t use it to write, or even draft up posts (honest…).
AI generated “content” is still, in my opinion, a bit like non-alcoholic beer. It looks the same, tastes broadly the same - but there’s, just something missing. It isn’t quite there yet.
But in the hunt for a nice closing statement, let’s do what every self-respecting writer does when they have run out of ideas and ask Chat-GPT what it thinks about it’s new competitor.
Oh, Chat-GPT, what are you like?
Comments