Slide contents

1. 1. People in a company of agents

   Employment at Anthropic · Alexander Polomodov

2. 2. What improved for people?

3. 3. Continuing the human-centered view

4. 4. Activity needs an outcome check

5. 5. Different data answer different questions

6. 6. Observation does not establish cause

7. 7. Agent hours exceed human hours

8. 8. Output rises across job functions

9. 9. Adoption is aligned to an event

10. 10. Engineers add more code after adoption

11. 11. Growth may precede adoption

12. 12. Quality proxies cover some consequences

13. 13. Sales are closer to business outcomes

14. 14. Junior share of technical hiring declines

15. 15. Higher AI-task complexity, slower junior headcount growth

16. 16. 424 hours estimates potential reading

17. 17. Managing agents becomes work of its own

18. 18. Human coordination remains

19. 19. Agents add a communication layer

20. 20. Agents need the current decision

21. 21. Technical hiring notes emphasize autonomy

22. 22. Design the loop, not every response

23. 23. Attention per accepted outcome

24. 24. Where does judgment come from?

25. 25. Who shares the gains from automation?

26. 26. Redeployment needs a real next role

27. 27. One workflow to test your assumptions

28. 28. Productivity includes people’s future