@GregKamradt: "Code and math are taking off because they are easy to verify, the next frontier is domains that are hard to verify" Th…
Summary
Greg Kamradt proposes a 7-level spectrum of verification difficulty for AI, ranging from instantly verifiable domains like math and code to civilization-scale systems with slow, noisy feedback.
View Cached Full Text
Cached at: 05/21/26, 03:47 PM
“Code and math are taking off because they are easy to verify, the next frontier is domains that are hard to verify”
This got me thinking - what does the spectrum of “easy to verify” look like?
This is loosely aligned w/ @DarioAmodei’s “intelligence bottlenecked” domains.
My take of easy > hard:
- Level 1: Instant, objective verification Math, code, formal proofs, chess tactics, parsing
AI improvement is easiest here because the loop is tight
- Level 2: Fast but incomplete verification Software engineering, UI implementation, data analysis, security bug finding
You can test a lot, but not everything. “It passes tests” is not the same as “it is good”
- Level 3: Human-evaluable creative work Copywriting, design, video thumbnails, sales emails, landing pages
Verification is possible through humans or markets, but noisy. AI can improve by predicting human reaction, but taste shifts and metrics can be gamed
There is no “right” answer, only feedback from humans
- Level 4: Market-verifiable work Startups, investing, product strategy, hiring, pricing, distribution
Reality gives feedback, but slowly and with tons of confounders
- Level 5: Experimentally verifiable science Materials, biology, chemistry, medicine, robotics
There is ground truth (physics), but experiments cost time and money. AI helps most when it can propose better candidates and reduce search space
- Level 6: Institutionally verifiable systems Education systems (Alpha school), legal systems, city planning, corporate management systems
You can measure outcomes, but the feedback cycle is long, and the counterfactual is hard
- Level 7: Civilization-scale verification Democracy variants, alternative governance, monetary systems, cultural norms, geopolitical strategy
Verification is slow, morally loaded, noisy, and often impossible to isolate. You may never get a clean answer, only accumulated historical evidence
How do you put a price on verifying a novel?
Are they though??
Math and code aren’t easy to verify? At least more so than writing a novel?
How do you standardize and formalize the process of writing a novel? Or creating a start up idea?
Similar Articles
@levie: We’re going to be in for a strange dynamic which is that some of the “hardest” work in the world is actually prone to a…
Levie and Max Spero discuss how verifiability determines which 'hard' knowledge work gets automated first, arguing that objectively testable domains like math, cyber, and code will see faster AI automation than fields with subjective outcomes.
@GergelyOrosz: There’s a popular theory that AI will finally make formal verification mainstream because mathematical proof of correct…
A podcast episode featuring Hillel Wayne discusses whether AI will drive mainstream adoption of formal verification, highlighting TLA+ use at Amazon and the challenges of writing formal specs.
What Happens When the World is Run on Code No One Understands?
The article discusses how AI is accelerating discoveries but creating a bottleneck in human verification, especially in critical software, and advocates for formal methods to ensure code correctness.
The Verification Horizon: No Silver Bullet for Coding Agent Rewards
This paper explores the challenges of verifying AI coding agents' outputs, arguing that verification is becoming harder than generation as models improve. It analyzes four reward constructions and shows that no fixed reward function remains effective as model capability grows.
@VitalikButerin: Many people have claimed that with AI-assisted bug finding, secure code (and hence trustless anything) will be impossib…
Vitalik Buterin shares an optimistic take on AI-assisted formal verification as a path to secure, trustless code, linking to his blog post explaining the basics of formal verification using Lean.