As AI commoditizes benchmarkable work, an organization's lasting moats lie in tasks that are verifiable through its private data and judgment
Sarah Guo
Context & Ripple Effects
Related coverage frames AI as compressing the value of work that can be evaluated against common benchmarks, while making it easier for companies to expose processes to AI systems. The counterweight is that trust, accountability, and domain-specific judgment remain important where outputs must be accurate and defensible.
This story sharpens that divide: durable differentiation shifts toward work whose quality can be checked using proprietary data and organizational judgment, rather than broadly reproducible model performance.
First-order effects
Organizations relying on standardized, easily benchmarked knowledge work face weaker differentiation as comparable AI capabilities become widely available.
Teams with proprietary data, established verification processes, and accountable expert judgment gain more leverage in deciding where AI can safely automate work and where it should remain supervised.
Second-order effects
Competitors will be pushed to compete less on generic AI-enabled output and more on access to distinctive data, workflow integration, and credible validation of results.
Buyers in regulated or high-stakes workflows are likely to place greater weight on provenance, reviewability, and responsibility for errors—not merely model capability—when selecting AI-enabled providers.
Third-order effects
If this pattern persists, AI markets may bifurcate between commoditized general-purpose production and higher-value systems embedded in private-data workflows with auditable human judgment.
The long-term moat shifts from possessing a model to governing data, verification, and accountability; firms that make every process fully machine-legible may also make more of their differentiation easier to copy.
The trend: AI is moving competitive advantage away from generic cognitive output and toward trusted, data-rich, verifiable workflows where organizations can stand behind the result.
the venture whitepill is that the world contains an uncountable number of opportunities that will remain inaccessible to the labs. however, recognizing this requires leaving (both physically and mentally) the city of San Francisco, and is thus out of reach for most investors
“So, we may ask two things of any kind of work. Is its correctness private and expensive to establish, the kind of truth that exists only inside someone's data? And is it walled off, locked inside a system you can't get into? Set those against how saturated the task is, and you
TL:dr on this excellent piece by @saranormous 1. Passing a test in school is not real life experience 2. Do hard and messy things 3. In the 90, investors were paralyzed because Microsoft would do everything. Now it is the foundation models. 4. Coding, which is the easiest
Good bit. This is very obviously not true to me. The level of groupthink psychosis amongst investors seems to be at an all time high. Why is that? [image]
Excellent essay. “The valuable work is illegible by construction: anything you can put on a leaderboard, you can train against, so anything measurable is already on its way to commodity.” Insights in every paragraph.
This is a critical post to read if you're building an applied AI company right now. “An application earns its place in the untrainable corner by doing unglamorous work: arranging a company's private reality so a model can act on it, handing the model the tools to act, working
Techmeme did what I could not with this essay, which was to cut through all the obviously Claude-generated portions to try to divine the author's original prompt [embedded post]