https://futurism.com/the-byte/ai-startups-money-trouble
AI foundation models are a commodity product and it’s unlikely their producers can sustain rentier status long enough to extract profit. This has been self-evident for years, indeed is inherent in the techno-optimism outlook from which they emerged. They don’t have network effect and brand loyalty or ever-accelerating advantages of scale that reify their position in the marketplace. They don’t have ever-accruing incumbent advantage.
So maybe build something more specific? Alas, no safe harbor there. AI solutions built in the usual manner of entrepreneurs “finding a problem and solving it well” are going to zero. Their moat / sustainable advantage won’t persist long enough for value extraction. Your company’s entire “solution” will become someone else’s feature will become someone else’s commodity afterthought before you even figure out how to hire enough people to ramp up. It’s hard to beat “free”.
PE ratios are based on the amount of time a business can hold the top of the hill and extract rent. This time will shorten.
On the other hand, rapid spikes in earnings will become more commonplace. Billion dollar (profit) flash in the pan will be the new norm. Winner take all. But very briefly.
The rise of tech saw us switch to a “crazy era” where “fundamentals” didn’t matter to valuation because it was a land grab, with the value all downstream in lifetime customer value, not present revenue streams. That won’t work anymore if the land you grab can’t be affordably defended and occupied long enough to matter. You’re disrupted before you even have a flag in the ground.
VC models, and the current institutionalization of hiring, employment, vesting, centralization of information flow and decisions, non-economic models of how internal resources are allocated, etc., can’t accommodate to this. They are rate-limited by human factors. Overhead for adaptation will be ever more of the fraction of time spent, until it hits a limit where it can’t happen fast enough (or if you prefer, accurately and efficiently enough). The “metadata” part of modern capitalistic institutional organization and process will be bloated bigger than the “data” part (the capacity to generate deliverable).
Something else is coming. We’ve known for about two years now. It will probably take longer than some think, but it will also happen more rapidly than some expect.
People don’t want drills. They want holes in the wall. Anything — physical products, organizations, processes, social constructs — that exists as an affordance or instrumentation for the purpose of information processing, smoothing, supply/demand discovery, information discovery, matching, storing, hoarding, facilitating, navigating, negotiating, knowing, judging, assessing, ascertaining, proxying, enabling, or anything else that is between wanting a hole in the wall and being a hole in the wall — will be disintermediated. They’re all going to zero.
Those functions will all happen. This isn’t a collapse. It’s a launch. It will be the end of most types of scarcity. There will be a virtuous flywheel spawning huge value disbursed throughout the commons. But it won’t have locally capturable monetized value. “Free” is not a business model.
Anticipating this, and knowing it will take quite a while yet, the strategy is to leapfrog. If you get far enough ahead on the value proposition, you gain some runway.
But you’re a train on a track with folks behind you trying to catch you. You add more and more fuel, to accelerate harder. But it’s asymptomatic. It’s $100 billion dollars to try to get some distance on those behind you. The advantage will go by the log of the effort, the resources, the expense.
And all the while, you’re facing the paradox that the best strategy for your would-be competitors is to do nothing.
Because if I have one year to catch you, there’s a good chance that my best resources for doing so will emerge in month 11. Try the leapfrog and bet on too shallow a jump and you arrive on Mars to find a thriving city there. Try to bet for too big a jump and you won’t get there at all.
That’s where we’re at now. The twelve-figure leapfrog. Its success would — will — put itself out of business.
Hang on!