Dominic Chiappe · People, capability & transformation

Thinking about how organisations perform in an AI-enabled world

The publications desk

Practical intelligence. Possible futures.

AyEye Today ·
A white filament stretches between violet masses above a saffron current sweeping through slate-grey waves.

Public capital enters. Control stays elsewhere

Anthropic’s IPO gives its benefit trust power over a board majority while a founder-owned entity retains 50.1 per cent of general voting power. Public investors gain economic exposure but little practical control, so the assurance question is whether the separate rights, triggers and escalation routes change difficult decisions in practice.

A company can sell you a piece of its future without selling you much say over it.

Anthropic priced its initial public offering on 18 November at $83 a share. The offering covers 980 million Class A shares, including 956,351,354 sold by Anthropic and 23,648,646 sold by existing holders. That puts about $81.3 billion of stock into the offer before any underwriter option. Trading is expected to begin on Nasdaq on 19 November under the symbol ANT.

Puzzle
Word search

Beyond the daily editions

In the conversation

All editions →
Follow a subject

Read around the story

AI, authority & assurance Workforce architecture Agent competency assurance
Go deeper

Ideas with a longer view

AyEye Papers AyEye Opinion · The Human Dividend HCM Market Stack
Explore the editions

The publications desk

Practical Impact Tech Tack Futures notebook All editions
Listen to AyEye · Recordings and articles (38)

AI narration. Only available recordings are listed.

Listen to unheard recordings →
  1. When AI speeds up AI, the review cycle must move too3:48
  2. The agent is answering the test it thinks it is taking2:40
  3. A completed task can still break its mandate2:46
  4. The web can now say “no training”. The terms may still say yes.2:16
  5. An alert is not a brake.3:07
  6. A human signature does not show who made the judgement2:29
  7. A cyber agent can work for days. Its mandate must stay current.3:03
  8. AI can fill the queue. Reality still has to decide2:46
  9. Anthropic has divided control. Accountability still needs an address2:40
  10. Britain has funded the nodes. The value lives in the hand-offs2:34
  11. The probability is not the policy2:41
  12. The room was secure. Is the conversation?2:50
  13. Your agent should not inherit all of your authority2:38
  14. AI can find the flaw. The hard part is moving the fix2:23
  15. Your HR agent can be accurate and still make the wrong promise2:38
  16. The latency budget is becoming an intelligence budget2:34
  17. AI's social contract is moving onto the cap table2:15
  18. AI productivity is becoming a queue at the human brain2:54
  19. The AI border has moved inside the company2:57
  20. AI safety now needs a version number2:33
  21. A job can be decomposed. Responsibility cannot.1:43
  22. The winning AI model is becoming a temporary condition1:34
  23. Autonomous HCM can consume capability faster than it creates it2:42
  24. AI governance has moved from rules to capability2:30
  25. Human differences and AI behaviour at work2:36
  26. The AI build-out has reached the credit committee3:11
  27. The AI test has become a public incident3:42
  28. The human work left behind is being turned into a score3:45
  29. The fake is learning to hire real people3:38
  30. High capability has become the default setting6:39
  31. WOAAS: Workforce Outcomes as a Service7:38
  32. Who gets to decide what your work means?7:45
  33. AI has turned proof into a queue6:28
  34. Your jobs are changing faster than your career system7:43
  35. The watermark is not the proof. The detector is.6:16
  36. AI is entering the laboratory before it has learned to run the control room6:23
  37. Safety is now competing with the model for compute6:07
  38. The next scarce workforce is the people between the boxes7:24