# mollick-roon-asi-consulting-forward-deployed-engineering-2026-05-10

## Veille

Ethan Mollick's (Wharton) consistency test: we'll know AI labs truly believe in ASI the day they dissolve their *Forward Deployed Engineering* (FDE) teams. Public debate with roon (OpenAI) on LinkedIn: roon objects that this is a **hayekian problem** (intelligence does not automatically resolve organizational information flow) and revives the term "**Gentle singularity**". Consensus in the comments: technology is the easy part; internal politics / legacy workflows / contractual liability are the real bottleneck. Marker phrase: *"Curing cancer might be easier than replacing Accenture"*. Epistemic **East Coast vs West Coast** opposition on the trajectory of AI adoption.

## Titre Article

You will know that the AI labs believe in ASI when [they dissolve their forward deployed engineering teams]

## Date

2026-05-10

## URL

https://www.linkedin.com/posts/emollick_you-will-know-that-the-ai-labs-believe-in-ugcPost-7460001935568510977-jXI0/

## Keywords

ASI (Artificial Super Intelligence), Forward Deployed Engineering (FDE), AI consulting, organizational change, hayekian problem, Gentle Singularity, enterprise transformation, AI adoption, organizational complexity, contractual liability, Accenture, consulting market, 2035 employment scenarios, white-collar automation, East Coast vs West Coast opposition, AI capability jaggedness, Wharton, OpenAI, sociology of technology adoption

## Authors

Ethan Mollick (professeur Wharton, auteur *Co-Intelligence*) — auteur du post ; roon (employé OpenAI, identité publique anonyme, voix influente du cercle accel) — interlocuteur cité ; commentateurs anonymes (praticiens, consultants, chercheurs).

## Ton

**Profile**: Mollick's voice (public intellectual, management professor at Wharton, prolific AI commentator on LinkedIn and his blog *One Useful Thing*), academic-popularized register with a sharp opinion. Long-form LinkedIn post format that blends a rhetorical *test* (a single sharp criterion to distinguish labs' stated belief from their actual behavior) with a public debate visibly preserved with roon. Target audience: AI practitioners, consultants, transformation executives, researchers in the sociology of technology adoption.

**Style**: Three devices. (1) **The single-criterion test** — "we'll know X when Y" — an elegant rhetorical figure that forces the other side to show its actions rather than its predictions. (2) **The concession that stings** — Mollick grants roon that the hayekian problem is real, then flips the argument: "your own FDE teams prove you know AI is not self-adopting, so how do you maintain the 2035 prediction?". (3) **The pithy formula** from the comments — "Curing cancer might be easier than replacing Accenture" — which condenses the shift of the debate from technical capability to organizational sociology. Overall tone is ironic and analytical, with the **East Coast / West Coast** rhetorical opposition presented not as geographic but epistemic. Mollick closes on a bilateral note, roon sums up with "Gentle singularity" — apparent agreement on the object, persistent disagreement on the timeline.

## Pense-betes

- **Mollick consistency test** (worth memorizing as a heuristic): a lab that *truly* believes in imminent ASI should be laying off its FDE and consulting teams, not hiring them. Otherwise it is either inconsistency, or an implicit admission that ASI will not deploy itself.
- **Hayekian problem**: Friedrich Hayek (*"The Use of Knowledge in Society"*, 1945) — useful information within an organization is **distributed**, tacit, contextual. Making it actionable requires human channels. More central intelligence does not solve this problem; aggregation protocols are needed.
- **Gentle singularity**: term coined by roon (OpenAI employee) — the transition to AGI/ASI will be **slow, fragmented, mediated**, not a *fast takeoff*. Sam Altman reuses the term in his essay *The Gentle Singularity* (June 2025) — an idea circulating from inside OpenAI to the outside.
- **East / West Coast opposition**: not geographic but epistemic. East = slow fragmented transformation + capability *jaggedness* + social constraints (Mollick). West = massive, fast automation once capabilities are reached (frontier labs).
- **Why Accenture survives** (the most counter-intuitive insight): not just competence but **contractual liability**. An AI cannot be sued. A consulting firm can. This asymmetry creates a persistent legal *moat* that models do not erode.
- **The canonical example**: *"ASI can produce the perfect transformation plan and still get stuck on a VP who refuses to modify their Salesforce workflow."* — the bottleneck migrates to the organizational level.
- **The lab contradiction** (worth putting to an executive over-investing in the ASI promise): *"If AI were truly self-adopting, why are your AI vendors selling FDE weeks per license?"*
- **Marker phrase**: *"Curing cancer might be easier than replacing Accenture"* — shareable.
- Companion reading with [[mollick-organizational-theory-agentic-ai-spans-control-2026-02]] (organizational theory of agentic AI: *spans of control*) — same author, same thesis.
- Implicit convergence with [[ng-the-batch-352-no-ai-jobpocalypse-2026-05-08]] (Andrew Ng, *"There Will Be No AI Jobpocalypse"*, same week) — anti-mass-disruption narrative.
- Convergence with [[wallace-wells-nyt-magazine-ai-populism-altman-backlash-no-one-ready-2026-05-08]]: labs sell a replacement narrative that runs into social and organizational reality.
- **Possible strategic cynicism**: FDE teams fund the long-term bet in the short term. Commercially rational, but it worsens the contradiction in the ASI narrative.
- **Direct link to Forward Deployed Engineering**: Palantir coined the term, OpenAI and Anthropic adopted it. Today AI labs' FDE roles are among the most dynamic tech job markets — which factually contradicts the self-adoption narrative.

## RésuméDe400mots

Ethan Mollick (Wharton) launches a consistency test aimed at frontier AI labs on LinkedIn: *"we'll know labs truly believe in ASI when they dissolve their Forward Deployed Engineering teams"*. As long as humans are needed to integrate AI into client organizations, white-collar jobs are not threatened in the short term.

**The debate with roon** (OpenAI employee, influential voice of the *accel* circle on X): roon retorts that this is a **hayekian problem**. Reference to Hayek (*"The Use of Knowledge in Society"*, 1945): useful information within an organization is tacit, distributed, contextual. Central intelligence, even superintelligent, does not automatically resolve information flows. roon says he is more optimistic about employment than the average lab precisely for this reason. Mollick concedes the hayekian point then turns the argument around: if AI is not self-adopting, then the lab prediction that "most white-collar jobs will be replaced by 2035" is contradicted by their own FDE teams. roon sums up the apparent agreement: *"Gentle singularity"* — the transition will be slow and mediated, not a *fast takeoff*. Sam Altman will later reuse this term in his essay *The Gentle Singularity* (June 2025).

**The consensus in the comments** (practitioners, consultants, researchers) converges on four points:

1. Technology is often the easy part. The real obstacle is internal politics, HR incentives, legacy systems, and above all the question "who is liable when it breaks?".
2. ASI can produce a perfect transformation plan and still get stuck on a VP who refuses to modify their Salesforce workflow.
3. Accenture (and consulting more broadly) survives because it sells **contractual liability**, not just competence. An AI cannot be sued. A firm can.
4. Most shared phrase: *"Curing cancer might be easier than replacing Accenture"* — a technical problem has clear success criteria; an organizational problem does not.

**The structuring tension** that Mollick formalizes: **East Coast vs West Coast** (epistemic, not geographic). East = slow, fragmented transformation, constrained by the *jaggedness* of capabilities and social complexity. West = fast, massive automation as soon as capabilities are sufficient.

**Conclusion**: labs sell ASI but hire consultants. This is either a logical contradiction, or — a more cynical view — a short-term revenue strategy that funds the long-term bet. Either way, their own FDE teams attest that AI is not (yet) self-adopting. The adoption bottleneck migrates from the technical to the organizational.

## GrapheDeConnaissance

- Ethan Mollick —affirme_que→ la dissolution des équipes FDE serait la preuve d'une croyance réelle des labos en l'ASI (AFFIRMATION, 0.96)
- Ethan Mollick —travaille_chez→ Wharton (ORGANISATION, 0.98)
- roon —travaille_chez→ OpenAI (ORGANISATION, 0.95)
- roon —affirme_que→ l'adoption de l'IA est un problème hayékien de flux d'information organisationnel (AFFIRMATION, 0.94)
- roon —a_créé→ Gentle Singularity (CONCEPT, 0.92)
- AI labs —utilise→ Forward Deployed Engineering (FDE) (METHODOLOGIE, 0.97)
- AI labs —prédit→ remplacement majoritaire des cols blancs d'ici 2035 (AFFIRMATION, 0.93)
- Accenture —est_basé_sur→ responsabilité contractuelle (CONCEPT, 0.93)
- Ethan Mollick —affirme_que→ l'ASI peut être bloquée par un VP refusant de modifier son workflow Salesforce (AFFIRMATION, 0.88)
- Ethan Mollick —soutient→ Vision East Coast (IA) (CONCEPT, 0.94)
- Ethan Mollick —s_oppose_à→ Vision West Coast (IA) (CONCEPT, 0.92)
- argument flux d'information organisationnel —s_inspire_de→ Friedrich Hayek (PERSONNE, 0.9)
- Palantir —a_créé→ Forward Deployed Engineering (FDE) (METHODOLOGIE, 0.85)
- Sam Altman —utilise→ Gentle Singularity (CONCEPT, 0.88)
- Praticiens et consultants —affirme_que→ la technologie est la partie facile (AFFIRMATION, 0.92)

---
Canonical: https://www.thekb.eu/en/fiches/mollick-roon-asi-consulting-forward-deployed-engineering-2026-05-10/