# sankar-vorflux-great-flattening-manifesto-2026-07-14

## Veille

Prasanna Sankar (co-founder/CTO of Rippling, founder of Vorflux) publishes "The Great Flattening" — a manifesto-essay arguing that coding models have become **superhuman** and that the bottleneck has shifted from code production to **encoding judgment** into *agent harnesses*. Everything inside the organization "collapses toward the harness"; everyone's real work becomes *self-profiling*: extracting the tacit decision frameworks from one's head to encode them into the codebase. Simultaneous launch of Vorflux ("autopilot for software engineering"), $15M seed (Y Combinator, Peak XV Partners, Alliance DAO). The essay drew 60,000+ views on X in 24 hours.

## Titre Article

The Great Flattening

## Date

2026-07-14

## URL

https://x.com/myprasanna/article/2077065557204222238

## Keywords

Great Flattening, Vorflux, Prasanna Sankar, Rippling, autopilot software engineering, agent harness, judgment vs code, self-profiling, organizational flattening, coding agents cloud, copilot vs autopilot, span of control, AI middle management, superhuman programming models, Y Combinator seed, Peak XV Partners, Alliance DAO, org chart flattening, bottleneck shift

## Authors

Prasanna Sankar (Prasanna S, @myprasanna) — co-fondateur et ex-CTO de Rippling ($16B+ valorisation), fondateur et CEO de Vorflux AI. Article publié sur X (format long-form article) le 14 juillet 2026, contenu repris comme manifeste Vorflux (vorflux.com/manifesto).

## Ton

**Profile:** Founding manifesto | Visionary-Provocateur | Techno-Philosophical | Expert-Insider

Sankar writes as a founder-engineer launching a product and a thesis at once — the manifesto is inseparable from the Vorflux announcement. The tone alternates between blunt technical assertion ("the models got superhuman at programming, genuinely superhuman right now") and organizational vision bordering on the philosophical ("self-profiling" as the new real work). Argumentative structure: state of affairs → bottleneck shift → organizational consequences → proposed solution (Vorflux). The writing is direct, without academic hedging, short sentences. The "copilot vs autopilot" framing serves both as an analytical lens and a product positioning — which is at once the essay's strength (memorable clarity) and its limit (evident founder bias). The parallel with previous waves (low-code, offshoring) that failed to deliver the promised flattening shows a historical awareness rare in this type of manifesto. Audience: founders, CTOs, early-stage investors, engineering leaders rethinking their org chart.

## Pense-betes

- **Central thesis**: the bottleneck in software engineering has shifted from code production to encoding judgment into agent harnesses
- **"The models got superhuman at programming"** — not trending, now
- **Copilot vs autopilot**: copilot requires staying at the controls; autopilot handles the entire route
- **"Everything inside the boundary collapses toward the harness"** — planning, design, architecture, review, execution
- **Self-profiling**: the real work becomes extracting tacit decision frameworks from one's head into the codebase
- **Vorflux**: a team of fresh agents, each with its own context, its own model, its own standing task
- **$15M seed** led by Y Combinator, Peak XV Partners, Alliance DAO
- **Reception**: 60,000+ views in 24h on X; Matt Shumer: "Vorflux is the single best coding agent I've ever used. It blows Devin out of the water."
- **Historical nuance**: the low-code and offshoring waves both promised organizational flattening — neither delivered at the predicted scale, and engineering headcount grew in both cases
- **Veille dossier convergence**: joins Cherny "coding is solved" (2026-05), Osmani harness engineering (2026-04-19), Trivedy Ralph Loop (2026-03-10), fortune.com "AI flattening corporate hierarchies" (2026-06-09)
- **The scarce skill**: knowing what to build (customer insight, product judgment), not how to build it
- **Org chart**: not shrinking so much as changing shape — from a hierarchical tree to a harness-centered network

## RésuméDe400mots

Prasanna Sankar, co-founder and former CTO of Rippling (valued at $16B+), publishes **"The Great Flattening"** — a manifesto-essay laying out a radical thesis on the future of software engineering. Published on X on July 14, 2026, simultaneously with the launch of Vorflux, his new startup building an **autopilot for software engineering**, funded by a **$15M** seed round led by Y Combinator with Peak XV Partners and Alliance DAO.

**Central thesis: the bottleneck shift**

Sankar argues that frontier coding models have crossed a threshold that most engineering teams have not yet recognized: **"the models got superhuman at programming — not trending there, genuinely superhuman right now."** As a result, the bottleneck has shifted from code production to **encoding judgment** into the *agent harnesses* that plan, test, review, and deploy that code. The scarce skill is no longer producing code, but **specifying and supervising the agents** that produce it. What remains is *customer insight* and *product judgment* — knowing **what** to build, not how.

**Collapse toward the harness**

The essay argues that **"everything inside the organizational boundary — planning, design, architecture, review, execution — collapses toward the harness,"** the system orchestrating the work. The org chart is not so much shrinking as it is **changing shape**. Everyone's real work becomes **self-profiling**: extracting the tacit decision frameworks from one's head to encode them into the codebase — "what decision-making framework is in your head that isn't in the codebase, how you triage, what data you reach for, the contrarian call nobody else would make."

**Copilot vs autopilot**

Sankar distinguishes the **copilot** model (staying at the controls, approving every turn) from the **autopilot** model (the agent handles the entire route). His thesis: the models are good enough for autopilot, but the tools haven't kept up. Vorflux proposes to address this with a **fresh-agents** architecture, each with its own context, its own model, and its own standing task, rather than a single giant session that drifts.

**Nuances and historical context**

The essay acknowledges that predictions of organizational flattening have a history of arriving prematurely — the **low-code** and **offshoring** waves of previous decades promised similar outcomes, yet **engineering headcount grew** through both. Reception has been massive: **60,000+ views on X in 24 hours**, with notable endorsements including Matt Shumer ("Vorflux is the single best coding agent I've ever used. It blows Devin out of the water.") and Sreeram Kannan ("coding agents on the cloud that scale infinitely"). The essay fits within the 2026 wave of manifestos on AI-driven managerial flattening, alongside Fortune (June 2026), Forbes, Fast Company, and Lepaya, but stands out for its **technical** grounding (the harness as the structuring unit) rather than a purely **organizational** one (middle management as the target).

## GrapheDeConnaissance

- Prasanna Sankar —publie→ The Great Flattening (DOCUMENT, 0.98)
- Prasanna Sankar —a_créé→ Vorflux (TECHNOLOGIE, 0.98)
- Prasanna Sankar —a_créé→ Rippling (TECHNOLOGIE, 0.98)
- Prasanna Sankar —affirme_que→ "the models got superhuman at programming, genuinely superhuman right now" (CITATION, 0.97)
- Prasanna Sankar —affirme_que→ le goulot d'étranglement a basculé du code vers l'encodage du jugement dans les agent harnesses (AFFIRMATION, 0.96)
- Vorflux —s_oppose_à→ modèle copilot (approbation manuelle à chaque tour) (CONCEPT, 0.93)
- Vorflux —permet→ autopilote logiciel avec agents frais à contexte propre (CONCEPT, 0.95)
- Y Combinator —soutient→ Vorflux (lead $15M seed) (TECHNOLOGIE, 0.97)
- Peak XV Partners —soutient→ Vorflux ($15M seed) (TECHNOLOGIE, 0.95)
- Alliance DAO —soutient→ Vorflux ($15M seed) (TECHNOLOGIE, 0.93)
- agent harness —améliore→ organisation ingénierie (planning, design, architecture, review, exécution) (CONCEPT, 0.94)
- self-profiling —s_applique_à→ ingénieurs extrayant leurs frameworks de décision tacites (CONCEPT, 0.93)
- modèles frontière codage —réduit→ importance de la production de code manuelle (CONCEPT, 0.95)
- Matt Shumer —affirme_que→ "Vorflux is the single best coding agent I've ever used. It blows Devin out of the water." (CITATION, 0.95)
- Prasanna Sankar —affirme_que→ les vagues low-code et offshoring n'ont pas livré l'aplatissement organisationnel promis (AFFIRMATION, 0.9)

---
Canonical: https://www.thekb.eu/en/fiches/sankar-vorflux-great-flattening-manifesto-2026-07-14/