“but am still waiting for evidence that supports anything close to a 10% chance of human extinction.”
Lambert challenges the evidentiary basis for an extinction probability near 10%, while acknowledging serious AI safety challenges. He argues that current evidence about AI improving training software and performing superhumanly in some domains does not establish near-term human extinction.
- Speaker
- Nathan Lambert
- Capacity
- Personal public post by Nathan Lambert; not established as an institutional statement.
- Observed
- September 10, 2026
- Published here
- September 10, 2026
- Classification rationale
- Challenges the evidence for the forecast, with explicit acknowledgment of safety concerns and later qualification of his rhetoric.
Surrounding material
Full context
Before the quotation
The post responds to the widely viewed Coxon warning and the accompanying extinction-risk discussion. A follow-up identifies the source of the 10% number; Hubinger had expressed it as his personal forecast.
After the quotation
Lambert subsequently accepted pushback that calling a sincerely held warning fearmongering may be unfair. He instead argued that company culture can drive beliefs unsupported by evidence. Preserve that qualification when describing his initial criticism.
Evidence
Complete source record
Full primary public post and available surrounding thread. Related context: https://x.com/natolambert/status/2097701194026127395 https://x.com/natolambert/status/2097703003750822333 https://x.com/natolambert/status/2097702300483473865 Speaker and position scope are explained in the linked event.