Schlep
The tedious, labor-intensive integration work required to make an AI model actually useful in a business workflow.
As defined in Leopold Aschenbrenner — 2027 AGI, China/US super-intelligence race, & the return of history
Other definitions
A tedious and unpleasant task, often a necessary part of building a business.
#276 Paul Graham’s Essays Part 2
Where it comes up · 6
- Modernizing Compliance7 claims
- Leopold Aschenbrenner — 2027 AGI, China/US super-intelligence race, & the return of history2 claims
- #190 Henry Ford and Thomas Edison1 claim
- #276 Paul Graham’s Essays Part 21 claim
- Leopold Aschenbrenner — 2027 AGI, China/US super-intelligence race, & the return of history1 claim
- Paul Christiano — Preventing an AI takeover1 claim
Most important claims · 10 of 13Search all claims for “Schlep” →
- Even a GPT-6-scale model (roughly five orders of magnitude effective training compute past GPT-4) has perhaps a 50% chance of still requiring large amounts of 'schlep' — data gathering, fine-tuning, changing workflows — before becoming a drop-in replacement for human labor, separate from whether it is intrinsically smart enough.Paul Christiano — Preventing an AI takeover
- Firestone saw that early car tires were so bad that rich owners hired a chauffeur-mechanic just to change them, but most people overlooked this painful problem, an example of schlep blindness, the unconscious dislike of tedious tasks that keeps people from seeing valuable opportunities.#190 Henry Ford and Thomas Edison
- Six states regulate sexual harassment training with requirements different enough that companies cannot use the same course nationwide, which is a source of schlep; a company that solves this mapping turns the pain into a competitive advantage.Modernizing Compliance
- Entering a complicated regulated space with fresh eyes can produce better solutions than insiders' acceptance of the existing system, because founders may not know how daunting the details are before starting.Modernizing Compliance
- Schleps are not merely inevitable but pretty much what business consists of; a company is defined by the tedious and unpleasant tasks it will undertake, so one should not shrink from them if they lead to something great.#276 Paul Graham’s Essays Part 2
- Entering a complicated regulatory space creates a moat when the startup absorbs the schlep of figuring out regulations, because customers who face a pain point value the solution once it is done.Modernizing Compliance
- Six states regulate harassment training with requirements different enough that identical training cannot be used across New York, California, and Connecticut, which Roxanne suspects reflects lobbying by the e-learning industry rather than any substantive difference in the content.Modernizing Compliance
- Large-scale engineering to make training runs work is tacit knowledge that China will eventually figure out, so it's not a durable source of advantage, but self-play RL for overcoming the data wall is more novel.Leopold Aschenbrenner — 2027 AGI, China/US super-intelligence race, & the return of history
- GPT wrapper companies are making a poor bet because they're betting on stagnation—assuming intermediate models will require extensive schlep to integrate—when actually the trajectory points to drop-in remote worker AGI that will make wrapper functionality obsolete, sonic boom-ing past the intermediate capability regime.Leopold Aschenbrenner — 2027 AGI, China/US super-intelligence race, & the return of history
- Much of AI's economic usefulness depends not on raw intelligence but on 'unhobbling'—unlocking capabilities like long-horizon agentic tasks and computer use; by 2027-2028, unhobbling will transform models from chatbots into drop-in remote workers that you interact with like coworkers (Zoom, Slack, assigning projects), which is far easier to integrate than intermediate models that require extensive workflow 'schlep'.Leopold Aschenbrenner — 2027 AGI, China/US super-intelligence race, & the return of history