資料來源#
- Anthropic's Boris Cherny: Why Coding Is Solved, and What Comes Next
- Designing with Claude: From prompt to production
摘要#
這是來自 Anthropic Labs 的 product,讓你用 plain language 描述想要什麼,並與 Claude 協作建立精緻的視覺 artifacts,包括設計、prototypes、slides、one-pagers 等等。它由一個「開發期間多數時候是三個人」的團隊打造,從想法到 launch「大約十週」,並在 Dan Carey 2026 年 5 月 talk 前約一個月以 research preview 上線(約 2026 年 4 月)。它是幾個 wiki concepts 背後的具體例子:Prototype Over PRD、Compounding Loop Optimization 與 Build for the Next Model。
起源#
團隊中的一位 designer Nate(先前是 Claude Code 的 designer)在投入另一個 bet 時,用一個週末 hack 出第一個 prototype。它刻意做得很薄:Agent SDK、一個非常薄的 IDE wrapper,以及他已經在 Claude Code 裡使用的一個現有 skill。第一個 prototype 是「一個 terminal 和一個 browser,差不多就這樣」。發布到 Slack 後,團隊的反應(「這裡有潛力 / 這裡完全壞掉,請修」)變成前幾週的 roadmap。Carey 的 framing 是:99% 的價值不是來自那個 prototype,而是來自「十週的 iterating and shipping,以及每天和使用者談話」。
關鍵功能(以及各自移除的 friction)#
多數功能一開始都是對團隊自己 workflow 的修補(Compounding Loop Optimization):
- Multiplayer — 對同一個 design 進行 real-time co-editing。打造它是因為 teammates 原本要一個接一個把 prototypes 來回交接;launch 後,這成為 users 的第一個要求。
- Handoff to Claude Code — 不必重新 export files,也不必重新輸入 conversation 的 context,就能把 design 推進 production。Launch 後 users 的第二個要求是:「那我要怎麼把這個放進 production?」
- Codebase exploration / GitHub linking — 用你的 brand,並建在你現有的 code 之上來 ship。
- Open export — designs 以 HTML, CSS, JavaScript export(不是封閉格式),反映「be an open tool」philosophy。
- Self-hosting — 團隊用「Claude Design to design Claude Design」,這是 Dogfooding as Product Discipline 最純粹的形式。
Product philosophy#
- Lift the floor, not the ceiling. 團隊打造過進階、pixel-level 的「power user」controls,少數 vocal testers 很喜歡,但「其他每個人都討厭」——令人困惑,而且「actively harmful」。他們在一週內把它們拔掉。教訓是:要做「a tool that lifts the level of craft for everybody, not just the ceiling on power users」(參見 Vibe Coding vs. Agentic Engineering 的「vibe coding raises the floor」)。
- Openness for needs you'll never meet. 因為有些 power users 想要 tool 不會支援的具體能力,所以 bet 是最大程度的 openness:HTML/CSS/JS export 加上 integrations,讓 specialized users 能把 designs 帶進自己的 tools。
- Forward-looking integration claim (verify). 截至 2026 年 5 月 talk,Carey 說「this week or next week」團隊會 ship 讓任何 design tool 透過它們現有的 MCPs 與 Claude Design integrate 的能力(見 MCP and Computer Use)。這被表述為即將發生,而非已確認 shipped;請把它當作帶日期的 forward claim,等待 verification。
Launch 與 iteration metrics#
- 從週五 launch 到接下來的週一之間 shipped 62 項 improvements,全部來自 launch-day user feedback(token efficiency、image handling、codebase exploration、near-instant exports)。Carey 說:這不是 all-nighter effort,而是「very natural」,因為團隊已經操練 daily loop 十週。
- 約一個月後將 token limits doubled,回應喜歡它、並想更常使用它的 users。
- 早期 prototype 的 capability gaps 不是由 engineering 補上,而是因 Opus 4.7 shipping 而關上——這是典型的 Build for the Next Model payoff。
在 wiki 中的先前提及#
Boris Cherny 曾把它標記為 vision example(錄於 Cowork):「Claude design I think is a really good example. It's pretty good today. It's going to get a lot [better].」Carey 的 talk 是第一個詳細說明它實際如何被打造的 account。
相關連結#
- Anthropic Labs — 孕育它的 incubator
- Dan Carey — 領導它並給出 build account 的 PM
- Claude Code — origin tech(Agent SDK、一個 Claude Code skill、薄 IDE wrapper)以及 handoff-to-production target
- Cowork — sibling Anthropic agent product(non-code knowledge work);Claude Design 是 visual-artifact sibling
- Anthropic — vendor
- Claude Opus 4.7 — 補上早期 prototype gaps 的 model release
- Prototype Over PRD — 快到能讓 prototypes 取代 PRDs 的 tool
- Compounding Loop Optimization — 它的 features 是結晶化的 loop-optimizations
- Build for the Next Model — 它最清楚的 case study
- Dogfooding as Product Discipline — 「use Claude Design to design Claude Design」
- Living Design System — HTML/CSS/JS export + design-system output 是同一種 portable-artifact move
- MCP and Computer Use — 「any design tool via MCP」的 integration surface
- Vibe Coding vs. Agentic Engineering — 「lift the floor for everyone」呼應 vibe-coding raising the floor
待解決的問題#
- 「any design tool via MCP」integration 是否真的在聲稱的 timeline 上 shipped?(來自 2026 年 5 月的 forward claim。)
- Claude Design 對 visual/aesthetic output 的 eval discipline 如何運作,在那裡沒有 compiler 或 test?(與 Cowork 針對 non-code artifacts 的 open question 相同;相關於 character/taste evals。)
資料來源#
- Designing with Claude: From prompt to production — Dan Carey,"Designing with Claude: From prompt to production," 2026 年 5 月
- Anthropic's Boris Cherny: Why Coding Is Solved, and What Comes Next — Boris Cherny 先前將它作為 vision example 提及
Cited by 17
- AI Native Product Cadence
Cat Wu's 6mo→1mo→1day cadence at Anthropic: research-preview branding, mission-as-tiebreaker, evergreen launch room, li…
- Anthropic
AI safety company / vendor of Claude; mission-as-tiebreaker culture; ~30–40 PMs across teams; Mike Krieger leads Labs r…
- Anthropic Labs
Anthropic's internal incubator — a 'bet factory' of ~a dozen tiny teams exploring the model frontier with lean-startup…
- Build for the Next Model
Prototype the thing that almost works, not the thing that already works: bet that the next concrete model release (not…
- Claude Code
Anthropic's agentic coding product; created by Boris Cherny late 2024; TypeScript/React; CLI/desktop/web/mobile/IDE sur…
- Claude Opus 4.7
GA frontier model from Anthropic; direct upgrade to 4.6 at same price; literal instruction following, 1.0–1.35× tokeniz…
- Compounding Loop Optimization
Dan Carey's discipline of instrumenting and automating every recurring step of the build loop — because when internal t…
- Cowork
Anthropic's non-code knowledge-work agent product; sibling to Claude Code; output is decks/inbox/dossiers; same MCP/com…
- Dan Carey
Product Manager leading product within Anthropic Labs; led Claude Design; 'Designing with Claude' talk (May 2026); ~two…
- Dogfooding as Product Discipline
Product sense is built by relentless first-hand use ("ant food"); Mr. Peanut catch; cross-source (Cat Wu vibe-checks, G…
- Engineer PM Convergence
Generalists across disciplines; product taste as bottleneck skill; Anthropic Claude Code team as case study; "just do t…
- Harness Shrinkage as Models Improve
Prompt scaffolding shrinks each model release; Cat Wu's pruning discipline; Boris Cherny "100 lines of code a year from…
- Entities — People, Orgs, Tools & Projects
Map of Content for all 32 entity pages. See Home for concept domains.
- Open Questions Backlog
_96 pages with open questions, as of 2026-06-14._
- Problem-Solution Fit Discipline
Idea-stage thesis: three defenses against premature building (time, resources, belief friction) all eroded; AI as devil…
- Product Velocity as Moat
Shipping speed as differentiator + trust signal ("you'll scale with us"); a treadmill that must convert into durable lo…
- Prototype Over PRD
Dan Carey's prototype-replaces-PRD method: record a why-not-what conversation, transcribe it, hand the transcript to Cl…
Related articles
- Cat Wu
Head of Product for Claude Code and Cowork at Anthropic; primary articulator of AI-native product cadence and engineer-…
- Claude Code
Anthropic's agentic coding product; created by Boris Cherny late 2024; TypeScript/React; CLI/desktop/web/mobile/IDE sur…
- Anthropic Labs
Anthropic's internal incubator — a 'bet factory' of ~a dozen tiny teams exploring the model frontier with lean-startup…
- Dan Carey
Product Manager leading product within Anthropic Labs; led Claude Design; 'Designing with Claude' talk (May 2026); ~two…
- AI Native Product Cadence
Cat Wu's 6mo→1mo→1day cadence at Anthropic: research-preview branding, mission-as-tiebreaker, evergreen launch room, li…
