Unverified50% confidenceFactExact time
There are mainly two ways to disprove ETH: one is to construct a concrete subexponential algorithm; the other is to prove through some non-constructive argument that such an algorithm must exist.
1
Sources
50%
Confidence
Long-term
Relevance
10/26/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedMethods for systematically constraining generative AI hallucinations include stripping out deterministic calculations, anchoring facts with verifiable external data sources, filtering exaggerations through adversarial secondary review, and requiring human approval71% similarUnverifiedYuri Gurevich puts forward the core argument: a true algorithm is always deterministic, and so-called "non-determinism" is merely the result of the algorithm obtaining information through an external "oracle".71% similarUnverifiedThe researchers argue that assuming reasoning traces possess human-like or algorithmically interpretable semantics is not only unnecessary but potentially misleading.68% similarUnverifiedFormal verification attempts to use mathematical proofs to guarantee the safety of AI behavior, but its scalability faces severe challenges.68% similarUnverifiedETH has not been proved or disproved to date, but it has become a cornerstone assumption in fields such as parameterized complexity and exact algorithm design.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1101465API
curl https://kongchang.com/api/v1/knowledge/claims/1101465MCP
get_claim(id=1101465)