Skip to the findings

FRONTIER THOUGHT / FINDINGS AND NEXT STEPS

THE AI CHOICES

Better AI does not automatically mean better work or stronger learning. The useful test is what a person can do better, with responsibility and correction still clear.

The discussion so far22 November 2026

The questions we take into the world

THE QUESTION WE STARTED WITH

Who owns the gains, and who carries the risk?

WHAT THE SOURCES TELL US

Looking at the whole job raises a harder question: who gets the time AI saves? Any decision to expand the service must count quality, difficult cases and all staff effort. It must also say which result would make us stop.

THE NEXT QUESTION

Will the organisation commit the gains to a specific improvement people can verify, and let workers and users challenge the result?

Use the same municipal process from Day 2. Ask the accountable service owner to sign off an objective for the public service and a review date, with staff and user input. Do not convert a generic productivity percentage into a staffing reduction target.

Organizing a discussion on this question?

Use these findings and sources as the starting point for a panel, interview, email exchange or small roundtable.

Coming next: A real AI gain: who gets the time back? ↓

What this means in practice

The Day 2 synthesis established that a faster step can coexist with a worse overall experience. The next decision is who has authority to call the change a success. The workplace evidence does not supply a universal target for Greece, and METR’s measurement revision warns against letting selected tasks stand for the whole service. Our proposed decision record therefore has three parts. First, the organisation states what public or customer benefit it seeks: fewer unresolved cases, a shorter wait for difficult requests or more time for personal support. Second, it records whose work makes that benefit possible, including review and repair. Third, it sets conditions for expansion, redesign or a pause before seeing the results. Worker and user representatives should be able to question those conditions. There is also a distribution choice that no benchmark can settle: saved time could support better service, training, reduced pressure or simply higher volume. The moderator should ask which use the organisation will actually commit to. The next brief can then distinguish a measured operational gain from a managerial choice about sharing it. We have prepared that decision structure; no Greek trial or agreement is being reported here.

What remains open: The actual gain and its distribution remain unknown. METR’s methodological revision reinforces the need to record selection, rather than treat a headline estimate as a local target.

What the sources say (4)

Dario Amodei

If AI can absorb company knowledge and be replicated, why should integrating it into the economy be harder than hiring people?

Model capability and deployment move on different curves. Even unusually fast adoption is constrained by enterprise rollouts and organisational decisions.

Position, paraphrased · 13 February 2026Read the original answer ↗

Satya Nadella

If Microsoft no longer has access to OpenAI models, what gives you confidence it can build and maintain a leading AI lab?

Agents need identities, observability, storage and tools. Infrastructure should support several model families because architecture and supplier advantages can change.

Position, paraphrased · 12 November 2025Read the original answer ↗

Brynjolfsson, Li and Raymond

How did access to an AI assistant change customer support work?

In one customer support firm, AI assistance increased resolved issues per hour by 15% on average; effects differed by experience. This is evidence about a particular workflow, not a universal productivity forecast.

Position, paraphrased · 4 February 2025Read the original answer ↗

METR research team

What happens to measurement when people avoid doing some tasks without AI?

METR reports that selective participation, task selection and concurrent agents make its newer productivity estimates unreliable. A local test should record which work is omitted as well as which work is timed.

Position, paraphrased · 24 February 2026Read the original answer ↗
Download the questions and evidence ↓
Watch the original published conversations
Dario Amodei

We are near the end of the exponential

dwarkesh.comOriginal recording and context ↗
Satya Nadella

How Microsoft is preparing for AGI

dwarkesh.comOriginal recording and context ↗
HOW THIS QUESTION MOVES FORWARD
  1. 01A concrete case

    Use the same municipal process from Day 2. Ask the accountable service owner to sign off an objective for the public service and a review date, with staff and user input. Do not convert a generic productivity percentage into a staffing reduction target.

  2. 02Evidence that could change the answer

    Signed objective, comparison covering all costs, decision thresholds agreed in advance and review date.

  3. 03A question we can answer

    A new example can change the answer. What would it tell us to ask next?

WHO CAN HELP ANSWER THIS?

The people closest to the problem.

Accountable service owner

Which benefit will you commit to, who receives it and which outcome would prevent expansion?

Signed objective, comparison covering all costs, decision thresholds agreed in advance and review date.

Worker representative and frontline colleague

Which new tasks, pressure or lost learning opportunities must be counted before you endorse this change?

Review and repair log, work allocation before and during the trial, training and escalation requests.

User panel including a difficult case user

Did the agreed benefit appear in an actual completed case, and what remains harder?

One de identified before and after journey and an accessible route to contest the decision.

Each perspective can fill a different gap in what we know.

Sources and what remains uncertain

Read the original work alongside the analysis. For the local case, we still need the evidence requested above.

  1. Dario AmodeiWe are near the end of the exponential ↗13 February 2026 · amodei-dwarkesh-20260213
  2. Satya NadellaHow Microsoft is preparing for AGI ↗12 November 2025 · nadella-dwarkesh-20251112
  3. Brynjolfsson, Li and RaymondGenerative AI at Work ↗4 February 2025 · qje-ai-at-work-20250204
  4. METR research teamWe are Changing our Developer Productivity Experiment Design ↗24 February 2026 · metr-uplift-design-20260224

NEXT DISCUSSION

23 November 2026 · 09:00 Athens time

A real AI gain: who gets the time back?

Before expanding one Greek AI service, what evidence from complete cases would show a benefit after checking and correction, and who should receive the saved time?

WHERE WE BEGIN

The model may improve faster than the organisation. Compare Amodei on adoption, Nadella on the operating environment and Mitsotakis on visible benefits to people, using the workplace studies as the evidence test.

THE PRACTICAL OUTCOME

One page comparing the service: complete case time, correction work, user completion, omitted cases, benefit recipient and a clear rule for deciding whether to stop or expand.

The published perspectives behind this question

A WORKING BRIEF, READY TO EDIT

Make this the next discussion.

The finding, original sources and evidence requests come with you. Review the question and choose the people who can clarify it.

Saved briefs are visible to authorized site editors. Use public material or material approved for that workspace.

Organizer sign in to save ↗