Frozen Competence: Where Human Value Survives
Models commoditize yesterday's competence; your value is using that cheap competence to make something new.
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 3
- Confidence
- 80%
A mental model for where durable human value lives as AI improves. Models ingest all prior human work and make yesterday's competence cheap to deploy anywhere — so it floods the world (everyone can make a landing page, write a tweet) and, because everyone uses the same models the default way, it all looks the same and becomes commoditized slop. Human value moves to taking that 'frozen human competence' and using it to make something new and interesting for a particular situation. Structurally, models always trail the humans pushing past them.
Origin
Dan Shipper's framing of why the 'AI job apocalypse' won't happen, explaining what new model drops actually do to the value of skills.
Core principles
- 01Models make yesterday's human competence cheap, so it commoditizes and stops being valuable
- 02Default, basic use of the same models produces sameness — recognizable slop
- 03Human value is recombining frozen competence into something new for a specific situation
- 04Models are structurally always trailing the people using them to push ahead
How to run it
- 1
See the commoditization coming
Recognize that whatever competence a new model absorbs will become cheap and ubiquitous, so any output produced the default way loses distinctiveness and value.
Pro tip Ask what everyone will now be able to do trivially — that's exactly the ground that's about to turn into slop.
Watch out Competing on the commoditized layer (generic landing pages, default tweets) is a race to the bottom because it all looks the same.
- 2
Use cheap competence as raw material, not the output
Treat the model's commoditized output as frozen competence to build on, then apply human taste and judgment to make something new and interesting for your specific situation.
Pro tip Creativity and standing out from the slop only grow more valuable as generation gets cheaper — lean into what looks distinctly non-default.
- 3
Keep pushing past the frontier
Whatever new thing you make will eventually get absorbed into the models — so keep moving to the next situation-specific, not-yet-done thing, staying just ahead of what's been commoditized.
Pro tip Deep, harder, more generative questions in your domain are where humans keep the edge as basics get automated.
Watch out Don't expect a permanent moat from any single innovation; the models catch up, which is why the practice has to be continuous.
In the wild
When coding models made everyone able to produce code, the result wasn't firing engineers — demand rose, because you need engineers to judge 'this is all slop, how should it actually go in our codebase' and turn cheap output into coherent product.
→ Automation created more human work at the judgment layer rather than eliminating the role.
AI design tools make interfaces that 'all look the same,' but designers with real taste use the same tools to build distinctive, novel interactions that stand out from the slop everyone else ships.
→ Their creativity becomes more valuable precisely because generic generation is cheap and everywhere.
Common mistakes
Competing at the commoditized default layer
Producing the same default output everyone else gets from the same models yields interchangeable slop with no distinct value — you lose on the exact ground AI just made cheap.
Assuming cheap generation means humans are obsolete
Commoditizing yesterday's competence actually opens room for humans to recombine it into new things and answer deeper questions; reading it as replacement misses where value relocated.
Is it for you?
Best for
Professionals in AI-exposed fields deciding where to invest their skills as generation becomes cheap and abundant
Not ideal for
Anyone wanting a concrete step-by-step process rather than an orienting mental model for value creation
From the transcript
“What models do in general is they make yesterday's human competence cheap. And so, it becomes commoditized. It's not valuable anymore.”
“What humans do is we sort of go in there and we're like, "Yeah, we have all this like frozen human competence from yesterday. How…”
“it's going to be more and more valuable to do to stand out from all the slop that people are shipping and launching all constantly”
From the episode
The AI paradox: More automation, more humans, more work
Dan Shipper