BREAKING
Anthropic: 60%+ Odds on Self-Improving AI
0
%
odds by 2028
0
x
more code shipped
Theory vs Reality on Coding Tasks
Exposure
94
Adoption
30
What the Labor Report Measures
1
O*NET tasks
↓
2
LLM capability
↓
3
Claude usage
Mixed Reactions to the Findings
Optimists
●
Coding speedups
●
No mass job loss yet
Skeptics
●
Displacement risk
●
Calls for verification
No Jobs Cliff Yet, But Watch 2028
AI NEWS BLITZ
Anthropic leaders put better than even odds on recursive self-improvement by 2028.