When a ceiling becomes a floor
How value shifts follow attention cycles
We are at the beginning of the AI revolution.
As a builder, I think (hope?) we have a few years' grace period to build products that bridge the gap between what the world looks like today and what it could look like with AI. It feels like today’s GOLD RUSH. What happens after that? Who knows.
Story time
A year ago, not very long after OpenAI released their API, I built a product called BloodExplain. You’d upload your blood test results, GPT would analyze them and write a clear explanation.
It started generating a tiny bit of revenue, and then I killed it (not a post-mortem, just to give some context)
At some point I thought about raising money, and reached out to an investor who passed because “people could just use ChatGPT.”
Which is essentially the case against GPT wrappers.
GPT-wrappers = last-mile solutions
A “GPT wrapper”, also known as a last-mile solution, is an application that acts as a thin layer on top of an LLM. It often offers functionality that a ChatGPT-like product could technically provide, but packaged as a specific experience for a specific use case. It has its own website, its own design, and an entire flow built around a particular pain point.
Some people talk shit about GPT wrappers. But I think they are actually the future.
The argument against wrappers is simple: In the short run, people will just use ChatGPT instead of your product. In the long run, AGI will do everything anyway - software is done. I can understand the long-term hypothesis - hard to say today, but in the short run, I don’t think that’s true. So let’s address it.
Why would users pay for your product if they can use ChatGPT to solve most of their problems?
Take BloodExplain as an example. ChatGPT could already explain your blood results - say that’s 80% of the value. But I designed a product specifically for that task and made sure everything made sense for our users: the prompt, the design, the flow, the experience, etc. - the remaining 20%.
So the question becomes: Why build a product if your starting point is capturing only 20% of the market?
When the ceiling becomes the floor
My take is that these 20% are going to soon become the new 80%.
We haven’t adapted yet. ChatGPT still feels like magic. This will be over soon (notes on attention). Once people get used to ChatGPT, what we currently perceive as the “80%” will simply become the baseline. Taken for granted.
Once society adapts (“sir when that clock strikes midnight”), the remaining 20% - specialization, context, workflow, taste - will expand and determine where value actually lives. And the value will move to tailored, use-case-specific solutions. What looks like a “last-mile solution” today will become a full marathon tomorrow.
“How quickly the world owes him something he knew existed only 10 seconds ago.”
Louis C.K.
People say things like: “Bro, it’s so over for designers.” That’s short-sighted.
Before photography, painters created most of the value by simply capturing reality. Interpretation was the smaller part. After cameras made capturing reality trivial, the value of painting shifted toward interpretation, style, and artistic expression. Our ceiling becomes the floor.
“There’s a tremendous amount of craftsmanship between a great idea and a great product.”
Steve Jobs