Why the next generation of breakout products will be both blisteringly fast and unmistakably crafted.

1. Setting the Stage

Generative AI has collapsed product‑development cycles from months to days, yet users complain that too much new software feels rushed and soulless. Two essays crystallise the tension:

• “Speed × AI” by VC firm NFX—a manifesto that says the only moat left is velocity.

• “Why is quality so rare?” by Linear—a rallying cry for a renaissance of craft.

Understanding these apparently opposing gospels is the first step toward synthesising them.

Sources

• NFX – “Speed × AI”, https://www.nfx.com/post/ai-speed

• Linear – “Why is quality so rare?”, https://linear.app/blog/why-is-quality-so-rare

2. The Velocity Gospel — NFX’s “Speed × AI”

James Currier argues that generative AI “ratchets the speed bar up 10×.” Founders who aren’t running 20–100 experiments a week will forfeit network effects, data advantages, and investor interest. NFX prescribes:

1. Experiment first. Treat every feature as a reversible test.

2. Borrow shamelessly. Copy what works; differentiate later.

3. Ego detox. AI doesn’t procrastinate—humans must learn to ship with the same detachment.

The message: in an AI arms‑race, inertia is fatal.

3. The Craft Renaissance — Linear’s “Why Is Quality So Rare?”

Karri Saarinen counters that the scarcest asset in 2025 isn’t code but care. Linear’s playbook:

1. Taste over metrics. Small maker teams trust intuition before dashboards.

2. Zero‑bugs policy. Anything broken is fixed inside seven days.

3. Internal‑only MVPs. Customers never touch half‑baked flows.

Quality, Saarinen claims, “creates gravity”—users become evangelists, CAC trends toward zero, and profitability arrives early.

4. Our Take — Speed and Craft as One Flywheel

Seen together, the essays describe two halves of a single engine:

AI removes the cost of iteration; craft converts that freed‑up time into long‑term love.

Fast loops give teams more shots on goal, while disciplined polish makes every hit stick. Data from a16z shows AI‑native startups now reach $2–4 M ARR in < 12 months—but only those that pair rapid deployment with trust‑worthy UX keep churn below 5 %. Lenny’s interviews with Linear confirm that “speed and quality are positively correlated in elite teams.”

Sources

• Andreessen Horowitz – “In Consumer AI, Momentum Is the Moat”,

• McKinsey – “How an AI‑enabled PDLC fuels innovation”,

• Lenny’s Newsletter – “Linear’s secret to building beloved B2B products”

5. Detailing the Quality–Velocity Flywheel

Tactical guard‑rails

1. Dual OKRs. Track lead‑time to production and a quality bar (e.g., NPS ≥ 50, defects < 0.1/kLOC).

2. Polish budget. Reserve 10–15 % of every sprint for craft debt—refactors, a11y fixes, empty states.

3. Automate the boring, review the nuanced. Let AI write tests; humans own edge‑cases and ethics.

4. Rotate a “Craft Custodian.” One team‑member per cycle has veto power on shippable polish.

6. Closing Thoughts

The coming decade belongs to high‑taste rapid iterators—teams that wield AI as jet fuel yet still grip the steering wheel with white‑knuckled intent. Move 10× faster so that you can spend 10× more of your reclaimed attention on what Linear calls “quality without a name.” When speed funds craft and craft sustains speed, the paradox disappears—and a formidable moat appears.

Move fast, polish hard.

1. Cursor in a sentence

Cursor aims to replace hand‑written code with "describe‑and‑build" workflows. Users already let the AI write 40–50 % of their committed lines; the north star is 100 %.

2. Why this interview matters

• Fastest‑growing dev tool in years – Cursor hit $100 M ARR in 20 months and is now valued at $9 B.

• Unique data moat – Millions of model calls each day reveal what code gets accepted, tweaked, or rejected.

• A preview of tomorrow's IDE – Truell explains why they shipped a stand‑alone editor instead of a plug‑in (more on that in §6).

3. From autocomplete to "auto-everything"

Cursor's mental roadmap for handing work from humans to machines.

Cursor tracks paid power users — devs who rely on the AI 4–5 days a week — as its North Star metric.

4. Two real bottlenecks Truell calls out

1. Context windows – Mature SaaS codebases can top 100 M tokens; even million‑token models struggle.

2. Continual learning – Agents need fresh tests, feature flags, and design tweaks hourly, not via occasional fine‑tunes.

Cursor is wiring runtime logs, design files, and chat threads straight into the model's feedback loop to tackle both.

5. Why "taste" still wins

Perfect code generation can't decide what feels right—layout nudges, UX pacing, or subtle logic. Tomorrow's engineer is a "logic designer."

6. Business lessons worth copying

→ Strategic bets that let a tiny team outrun bigger rivals.

💡 Cursor still runs first‑round coding screens without AI tools; they'd rather train great engineers in AI than screen them out unfairly.

7. Action playbook for product teams

Here are a few things Cursor does, that you could copy too!

1. Instrument everything. Capture every user action as labeled data.

2. Write specs in plain English first. Let an agent draft code, then study the failure cases.

3. Run "taste reviews." Pair code reviews with short sessions judging clarity, delight, and ethics.

4. Plan for multimodal context. Pipe logs, screenshots, and Figma frames into one agent workspace.

5. Onboard juniors as agent supervisors. Teach reasoning and prompt craft before deep syntax.

8. Founding story & mindset

• Pivot from CAD to code – The team's first project was a CAD co‑pilot. They shelved it when they realised code had orders‑of‑magnitude more data and clearer scaling laws.

• "Follow the line" mantra – They bet that language models would keep improving on a predictable curve and built for the capability ahead, not the one in hand.

• Year‑long "wilderness" – Cursor spent ~12 months iterating with tiny user numbers before the product finally clicked and growth took off.

9. Moats in the AI era

Truell argues that distribution is the data moat: the more real‑world edits your model sees, the faster it learns — just like Google Search in 1999.

Closing thought

Programming is shifting from typing to directing. Cursor shows that the winners won't bolt models onto old tools; they'll rethink the tool itself and double‑down on the one thing machines still lack: human judgment.