A bear case for the AI industry, as the sectors LLMs seem capable of disrupting so far, like writing, digital art, and programming help, are not very lucrative
Gemini and the supply paradox of AI — Another day, another huge new AI model revealed. This time it's Google's Gemini.
The commercial question has become more consequential as later coverage reported diminishing returns from increasingly costly model-building efforts. The issue is not whether LLMs are useful, but whether the early use cases can support the cost of creating and operating frontier models.
First-order effects
The analysis pressures Google and other model developers to justify Gemini investment through revenue sources beyond low-priced writing, image-generation, and coding-assistance tools.
Businesses in those early-adopting categories face greater pressure to treat LLMs as productivity features or bundles rather than assume standalone AI products can command high margins.
Second-order effects
Model providers are pushed toward distribution-heavy surfaces and higher-value workflows; Google’s later AI Mode rollout in Search illustrates the appeal of placing AI inside an existing, high-reach product.
Competition shifts from announcing broadly comparable models to proving differentiated outputs, integrations, and willingness to pay, potentially making raw capability gains less commercially decisive.
Third-order effects
If model costs remain high while output in digital-content markets becomes abundant, AI value may accrue more to platforms with distribution and customer relationships than to standalone generation tools.
The pattern points to a synthetic-supply paradox: expanding model output can reduce the scarcity that supports pricing, so durable AI economics may depend on applications that measurably improve higher-value work.
The trend: Frontier AI is moving from a race to demonstrate model capability toward a test of whether distribution and workflow integration can convert abundant generated output into durable revenue.
This explains why economic disruption is low despite the success of the tech... because you can only automate adding to data sets, like writing, that are already massively oversupplied. Why? Because AIs basically *just* scale off of the supply-size of the data set! [image]
“Call it the supply paradox of AI: the easier it is to train an AI to do something, the less economically valuable that thing is. After all, the huge supply of the thing is how the AI got so good in the first place.”
Examples: AIs are good at writing Reddit comments, but producing Reddit comments is worth almost nothing. AIs are good at generating digital art, but the market for digital art is tiny. AIs are good at writing essays, but it's extremely hard to make a living writing essays.
I haven't found much use for AI in full-on writing large bodies for code for me, but I continue to be astounded by its ability to be a superb pair programmer. It knows all the APIs, we never need to Google anything, and its suggestions are often delightful. A+.
An interesting take by @erikphoel on whether AI is about to take all our jobs (and moneys). In short: The things it's best at, at least for now, aren't lucrative. https://open.substack.com/...
Spot on. “Call it the supply paradox of AI: the easier it is to train an AI to do something, the less economically valuable that thing is. After all, the huge supply of the thing is how the AI got so good in the first place.” https://www.theintrinsicperspective.com / ...
Is generative AI doomed to disrupt only low-revenue industries, like writing and digital art? My latest on the supply paradox: AIs get good because of massive data sets... but adding to the data set isn't lucrative, as there was already so much of it! https://www.theintrinsicpers…