OpenAI and Anthropic have each said that their AI models have disproven the Exponential Time Hypothesis (ETH), a long-standing conjecture in theoretical computer science, a claim that, if confirmed, would mark a significant theoretical breakthrough with possible consequences for how researchers assess encryption security.
AI MODELS · THEORETICAL COMPUTER SCIENCE
Two AI labs claim their models disproved a foundational complexity conjecture
OpenAI and Anthropic each say their models have disproven the Exponential Time Hypothesis (ETH). If confirmed, it would be a major theoretical breakthrough with possible implications for how encryption security is assessed. Neither the proof nor the technical details have been published yet.
2
AI companies making the claim: OpenAI and Anthropic
0
Proofs or technical details included in the report
0
Named researchers or confirmed outside verification
From claim to established result
The claim still has to pass each stage before it can be treated as a finding.
STEP 1
Company claim
Stated by OpenAI and Anthropic
→
STEP 2
Proof published
Full technical details released
→
STEP 3
Expert scrutiny
Mathematicians check the proof
→
STEP 4
Community verdict
Result accepted or rejected
If the claim holds
Some problems long considered intractable would admit faster algorithms.
Many results built on ETH would need to be reassessed.
Cryptographic hardness assumptions tied to ETH could come under review.
What caution requires
A theoretical result does not by itself break any deployed encryption system.
Practical impact depends on the proof's specifics and on cryptographers' follow-up analysis.
Until checked, this is a company-reported claim, not an established finding.
WHAT THE HYPOTHESIS SAYS
ETH holds that certain hard problems cannot be solved much faster than exhaustive search as they grow.
Many results across complexity theory assume ETH is true. Disproving it would mean those hard problems admit faster algorithms than previously believed, which is why the claim is being watched so closely.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article...