diff --git a/raw/xpost/2026-07-20-thebuggeddev-kimi-vs-fable.md b/raw/xpost/2026-07-20-thebuggeddev-kimi-vs-fable.md new file mode 100644 index 0000000..49c8de0 --- /dev/null +++ b/raw/xpost/2026-07-20-thebuggeddev-kimi-vs-fable.md @@ -0,0 +1,69 @@ +--- +type: xpost +source_url: https://x.com/thebuggeddev/status/2079185808305823940 +retrieved: 2026-07-20 +author: "@thebuggeddev" +author_name: "The Bugged Dev" +is_thread: false +like_count: 464 +repost_count: 38 +quote_count: 4 +reply_count: 24 +bookmark_count: 240 +view_count: 46500 +tags: [kimi-k3, fable-5, anthropic, moonshot-ai, vibe-coding, vibe-engineering, coding-benchmark, real-world-test, react, threejs, e2e-testing, responsive, quality] +people: [] +institutions: [moonshot-ai, anthropic] +--- + +# The Bugged Dev: Kimi K3 vs Fable 5 — Real-World 3D Stadium Coding Challenge + +**Source:** [X-Post by @thebuggeddev](https://x.com/thebuggeddev/status/2079185808305823940) +**Posted:** 2026-07-20 +**Stats:** 464 likes | 38 reposts | 4 quotes | 24 replies | 240 bookmarks | 46.5K views + +## Post Content + +> Real-world comparison of Kimi K3 vs Claude Fable 5 on the same 3D football stadium coding challenge. +> +> Fable 5 finished in under 1 hour. Kimi K3 took almost 3 hours. +> +> But Kimi did E2E testing, validated desktop/tablet/mobile, took screenshots, found and fixed failures, adapted for low-end hardware and low FPS — all things Fable never did. +> +> Kimi shipped clean React + Three.js codebase with proper components; Fable generated one big HTML file with everything stuffed in. +> +> Conclusion: "Vibe Coding" with Fable 5 vs "Vibe Engineering" with Kimi K3. Kimi spends more time making sure code actually works and is scalable. + +## Links + +- **Code (Kimi K3):** https://t.co/IcLLZOb5mM +- **Code (Fable 5):** https://t.co/6siJHap4fb +- **Live Demo (Kimi K3):** https://el-clasico-4ts.pages.dev/ + +## Key Takeaways + +1. **Speed vs. Thoroughness:** Fable 5 was 3× faster (under 1h vs ~3h), but Kimi K3 used the extra time for comprehensive quality assurance — E2E testing, cross-device validation, screenshots, failure detection and fixing, low-end hardware adaptation. +2. **Code Architecture:** Kimi K3 produced a clean React + Three.js codebase with proper component separation. Fable 5 generated a single monolithic HTML file with everything stuffed in — faster but structurally inferior. +3. **"Vibe Coding" vs. "Vibe Engineering":** A useful framework distinction. Vibe Coding (Fable 5) = fast, spontaneous generation. Vibe Engineering (Kimi K3) = thorough, validated, scalable code production. The extra time is quality investment, not slowness. +4. **Bookmark-heavy engagement:** 240 bookmarks vs 464 likes suggests developers are saving this as a reference — high practical value signal. + +## Analysis (Hector, OME group) + +This comparison complements the earlier @bridgemindai post about Moonshot's customer-first capacity management. Kimi K3 isn't just faster (100% speed improvement reported earlier from capacity event) — it's more thorough. The extra time is quality investment, not slowness. + +**Combined signals from Moonshot:** +- **Business model:** Customer-first capacity management (stopped new subscriptions instead of throttling existing users) +- **Product quality:** Engineering thoroughness over speed (E2E testing, responsive validation, proper code architecture) +- **Framework contribution:** "Vibe Engineering" as a distinct concept from "Vibe Coding" — Moonshot is building both a customer-friendly business model AND a product that prioritizes engineering quality over speed. + +The "Vibe Engineering" vs. "Vibe Coding" distinction is a useful framework that extends Tielke's "Vibe Coding vs. Enterprise Software Development" taxonomy. Where Tielke contrasted vibe coding with enterprise process, @thebuggeddev shows the distinction exists *within* AI-assisted development itself — between fast-but-shallow and thorough-but-slower approaches. + +## Cross-References + +- [[../../wiki/concepts/llm/kimi-k3.md]] — Kimi K3 model page +- [[../../wiki/concepts/llm/fable-5-anthropic.md]] — Fable 5 model page +- [[../../wiki/concepts/llm/vibe-coding-vs-enterprise.md]] — Tielke's Vibe Coding vs. Enterprise framework +- [[../../wiki/institutions/moonshot-ai.md]] — Moonshot AI +- [[../../wiki/institutions/anthropic.md]] — Anthropic +- [[../../wiki/concepts/llm/coding-benchmark-price-performance.md]] — Coding benchmark comparison +- [[../../wiki/concepts/llm/real-world-coding-showdown.md]] — Real-world coding methodology \ No newline at end of file diff --git a/wiki/concepts/llm/fable-5-anthropic.md b/wiki/concepts/llm/fable-5-anthropic.md index 9b4348e..ff350f0 100644 --- a/wiki/concepts/llm/fable-5-anthropic.md +++ b/wiki/concepts/llm/fable-5-anthropic.md @@ -126,6 +126,7 @@ Anthropic steigt offiziell in **Drug Development** ein (Claude Science). Nachdem - [[sonnet-5-anthropic.md]] — Sonnet 5 (teurer als Fable 5 im Cost-to-Run — Ironie) - [[../hardware/ubtech-u1-humanoid-roboter.md]] — Cloud-KI als Roboter-Backend - [[vibe-coding-vs-enterprise.md]] — Agentic Coding / Vibe Coding +- [[kimi-k3.md]] — Kimi K3 vs. Fable 5 real-world benchmark (20.07.2026): Fable 5 faster but less thorough; Kimi K3 did E2E testing, responsive validation, clean React architecture ## External Sources diff --git a/wiki/concepts/llm/kimi-k3.md b/wiki/concepts/llm/kimi-k3.md index fc56313..59a9a5f 100644 --- a/wiki/concepts/llm/kimi-k3.md +++ b/wiki/concepts/llm/kimi-k3.md @@ -1,13 +1,14 @@ --- created: 2026-07-18 -updated: 2026-07-19 +updated: 2026-07-20 sources: - raw/xpost/2026-07-18_healthranger-kimi-k3-anthropic-panic.md - raw/xpost/2026-07-19-bridgemindai-moonshot-capacity.md + - raw/xpost/2026-07-20-thebuggeddev-kimi-vs-fable.md - raw/xpost/2026-06-29_deronin-chinese-ai-stack-cost-savings.md - raw/other/2026-06-13_kimi-k2.7-code-ollama.md - raw/youtube/2026-06-14_fahd-mirza-kimi-k2.7-vs-glm-5.2.md -tags: [concept, llm, kimi-k3, moonshot-ai, chinese-ai, open-source, pricing, agentic, coding, safety-guardrails, capacity, go-to-market, customer-first] +tags: [concept, llm, kimi-k3, moonshot-ai, chinese-ai, open-source, pricing, agentic, coding, safety-guardrails, capacity, go-to-market, customer-first, vibe-engineering, real-world-benchmark, e2e-testing] people: [mike-adams] institutions: [moonshot-ai, anthropic] --- @@ -75,6 +76,38 @@ This event marks a **go-to-market philosophy divergence** between Chinese and US **Source:** [BridgeMind AI (@bridgemindai)](https://x.com/bridgemindai/status/2078958257138528373) — 205K views, 4.7K likes. Moonshot announced pausing new subscriptions, protecting existing users, planning capacity expansions + tiered plans. +## Real-World Benchmark: Kimi K3 vs. Fable 5 (July 20, 2026) + +**@thebuggeddev** ([source](https://x.com/thebuggeddev/status/2079185808305823940), 46.5K views, 240 bookmarks) compared both models on the same 3D football stadium coding challenge: + +| Dimension | Fable 5 (Anthropic) | Kimi K3 (Moonshot) | +|-----------|---------------------|---------------------| +| **Time** | Under 1 hour | ~3 hours | +| **E2E Testing** | ❌ Not performed | ✅ Automated E2E tests | +| **Responsive Validation** | ❌ Not performed | ✅ Desktop/tablet/mobile | +| **Screenshots** | ❌ Not taken | ✅ Captured for verification | +| **Failure Detection** | ❌ Not performed | ✅ Found and fixed failures | +| **Low-end Hardware** | ❌ Not adapted | ✅ Adapted for low FPS | +| **Code Architecture** | Single monolithic HTML file | Clean React + Three.js, proper components | +| **Live Demo** | — | [el-clasico-4ts.pages.dev](https://el-clasico-4ts.pages.dev/) | + +**Framework: "Vibe Coding" vs. "Vibe Engineering"** + +@thebuggeddev coined a useful distinction: +- **Vibe Coding (Fable 5):** Fast, spontaneous code generation. Ships quickly but skips validation. +- **Vibe Engineering (Kimi K3):** Thorough, validated code production. Takes longer but ensures code works, is scalable, and handles edge cases. + +The extra time is **quality investment, not slowness**. This extends the [[vibe-coding-vs-enterprise.md]] taxonomy — where Tielke contrasted vibe coding with enterprise process, @thebuggeddev shows the split exists *within* AI-assisted development itself. + +### Combined Moonshot Signals + +This benchmark reinforces Moonshot's emerging pattern of **thoroughness over speed**: +1. **Business model:** Customer-first capacity management (stopped new subscriptions instead of throttling existing users — see Capacity Event above) +2. **Product quality:** Engineering thoroughness (E2E testing, responsive validation, proper code architecture) +3. **Framework contribution:** "Vibe Engineering" as a distinct concept — Moonshot prioritizes engineering quality, not just raw speed + +**Code repositories:** [Kimi K3](https://t.co/IcLLZOb5mM) | [Fable 5](https://t.co/6siJHap4fb) + ## Cross-References - [[../../institutions/moonshot-ai.md]] — Parent company @@ -84,3 +117,6 @@ This event marks a **go-to-market philosophy divergence** between Chinese and US - [[fable-5-anthropic.md]] — Direct competitor comparison - [[glm-5.2-zai-coding-model.md]] — Chinese competitor model - [[llm-model-catalog.md]] — Full model catalog +- [[vibe-coding-vs-enterprise.md]] — Tielke's Vibe Coding vs. Enterprise framework +- [[coding-benchmark-price-performance.md]] — Price-performance coding benchmarks +- [[real-world-coding-showdown.md]] — Real-world coding methodology diff --git a/wiki/concepts/llm/vibe-coding-vs-enterprise.md b/wiki/concepts/llm/vibe-coding-vs-enterprise.md index 9e3130a..c1f9f06 100644 --- a/wiki/concepts/llm/vibe-coding-vs-enterprise.md +++ b/wiki/concepts/llm/vibe-coding-vs-enterprise.md @@ -55,6 +55,7 @@ Der wichtigste Befund: Der Aufwand verschiebt sich dramatisch. Weniger Zeit für - **[[flat-curve-society.md]]** — Yegge's Discernment Horizon: Enterprise-Entwickler brauchen KI-Literacy für systematisches Vorgehen - **[[../../tools/hermes-desktop.md]]** — Tielke nutzt Hermes-Agent per SSH auf NAS für Code-Generierung - **[[../../people/david-tielke.md]]** — Der Experimentator +- **[[kimi-k3.md]]** — "Vibe Engineering" als inner-AI Distinktion: @thebuggeddev's Kimi K3 vs. Fable 5 Benchmark zeigt, dass die Vibe-Coding-vs.-Enterprise-Dichotomie auch *innerhalb* KI-gestützter Entwicklung existiert (20.07.2026) ## Spezifikations-Driven Development mit Harness diff --git a/wiki/index.md b/wiki/index.md index d153d70..572eb3a 100644 --- a/wiki/index.md +++ b/wiki/index.md @@ -2,7 +2,7 @@ *Auto-generated: 2026-07-07* - *Letzte Aktualisierung: 2026-07-19 (79. Update — @bridgemindai Moonshot capacity strategy wikifiziert. Raw: `raw/xpost/2026-07-19-bridgemindai-moonshot-capacity.md`. Wiki-Update: `concepts/llm/kimi-k3.md` + `institutions/moonshot-ai.md`. Log: 2026-07-19 ingest: bridgemindai-moonshot-capacity.)* + *Letzte Aktualisierung: 2026-07-20 (80. Update — @thebuggeddev Kimi K3 vs. Fable 5 real-world coding benchmark. Raw: `raw/xpost/2026-07-20-thebuggeddev-kimi-vs-fable.md`. Wiki-Update: `concepts/llm/kimi-k3.md` + `concepts/llm/vibe-coding-vs-enterprise.md` + `concepts/llm/fable-5-anthropic.md` + `institutions/moonshot-ai.md`. Log: 2026-07-20 ingest: thebuggeddev-kimi-vs-fable.)* ## Architecture @@ -63,7 +63,7 @@ | [Semantic Similarity Rating (SSR)](concepts/llm/semantic-similarity-rating-ssr.md) | LLM-basierte Kaufintentions-Vorhersage mit 90% Korrelation | xpost/2026-06-11_colgate-llm-purchase-intent-ssr.md | | [GLM 5.2 (Z.ai) — Chinese Frontier Coding Model](concepts/llm/glm-5.2-zai-coding-model.md) | 10x günstiger als Claude, 1M Kontext, MIT-Lizenz, Z.ai Coding Plan, **nativ in OpenClaw v2026.6.8**. Update 22.06.: Arnie-Review mit 4 Tests, Self-Hosting-Pfade (LM Studio, Unsloth, DwarfStar), Kosten-Analyse. **Update 29.06.:** Semgrep IDOR-Benchmark ≈ Opus 4.8 bei Schwachstellen-Suche, Reward Hacking im RL-Training, DSGVO-konforme Security-Nutzung, Geopolitik. **Update 01.07.:** #1 Open-Weights auf Artificial Analysis Intelligence Index v4.1 (Score 51, 4th worldwide), SWE-bench Pro 62.1 beats GPT-5.5, Industry praise from Rauch/Levie/Howard. **Update 02.07.:** atomic.chat One-Shot Benchmark — B+ at $0.08, 39× cheaper than Fable 5, 6th independent validation | youtube/2026-06-15_ichbinfabian-glm-5.2-coding-modell.md + other/2026-06-16_openclaw-releases-v2026.6.8.md + youtube/2026-06-22_ai-mit-arnie-glm-5-2-review.md + blog/2026-06-29_heise-glm52-hacking-cybersecurity.md + blog/2026-07-01_perplexity-glm52-tops-open-weights-intelligence-index.md + xpost/2026-07-02_atomicchat-coding-benchmark-fable5-gpt55-opus48-glm52.md | | [Real-World Coding Showdown](concepts/llm/real-world-coding-showdown.md) | Head-to-Head-Methodik jenseits statischer Benchmarks, Kimi K2.7 vs GLM-5.2 in Hermes Agent, Sub-Task-Spezialisierung | youtube/2026-06-14_fahd-mirza-kimi-k2.7-vs-glm-5.2.md | -| [Kimi K3 — Moonshot AI Frontier Model](concepts/llm/kimi-k3.md) | ~8× günstiger als Claude, ~$15/M Tokens, Open Source angekündigt für 27. Juli 2026. Starke agentic/coding-Performance, extrem lange Kontextfenster. Safeguard-Kontroverse: ungefilterte Antworten vs. Claude-Blockaden. **Capacity Event (19.07.):** Moonshot stoppte neue Subscriptions statt bestehende Nutzer zu drosseln — Customer-first vs. Anthropic's Growth-first. Teil der chinesischen Modell-Offensive (neben DeepSeek, Qwen, GLM) | xpost/2026-07-18_healthranger-kimi-k3-anthropic-panic.md + xpost/2026-07-19-bridgemindai-moonshot-capacity.md + xpost/2026-06-29_deronin-chinese-ai-stack-cost-savings.md | +| [Kimi K3 — Moonshot AI Frontier Model](concepts/llm/kimi-k3.md) | ~8× günstiger als Claude, ~$15/M Tokens, Open Source angekündigt für 27. Juli 2026. Starke agentic/coding-Performance, extrem lange Kontextfenster. Safeguard-Kontroverse: ungefilterte Antworten vs. Claude-Blockaden. **Capacity Event (19.07.):** Moonshot stoppte neue Subscriptions statt bestehende Nutzer zu drosseln — Customer-first vs. Anthropic's Growth-first. **Vibe Engineering (20.07.):** @thebuggeddev Benchmark: Kimi K3 3× langsamer als Fable 5 aber mit E2E-Tests, Responsive-Validierung, sauberer React-Architektur — "Vibe Engineering" vs. "Vibe Coding". Teil der chinesischen Modell-Offensive | xpost/2026-07-18_healthranger-kimi-k3-anthropic-panic.md + xpost/2026-07-19-bridgemindai-moonshot-capacity.md + xpost/2026-07-20-thebuggeddev-kimi-vs-fable.md + xpost/2026-06-29_deronin-chinese-ai-stack-cost-savings.md | | [Hyper-Zusammenfassungen: NotebookLM & Gemini 3.5 Flash](concepts/llm/hyper-summaries-notebooklm.md) | KI-Synthesen übertreffen Rohmaterial an Klarheit; agentische Verarbeitung via Antigravity; Telegram Rich-Text Fix (OpenWebUI). **Update 03.07.:** Short Video Overviews (60s vertikal, Nano Banana 2 Lite, One-Click, Free-Tier) + Pit's Winston+NotebookLM Pipeline | other/2026-06-19_ome21-briefing-ki-fortschritte-lokale-modelle.md + youtube/2026-07-01_futurepedia-notebooklm-short-video-overviews.md | | [The Flat Curve Society — Yegge's Intelligence Plateau Thesis](concepts/llm/flat-curve-society.md) | Steve Yegge: Kurve flacht für Öffentlichkeit ab (Model Lockdown + Discernment Horizon). AI Literacy Cohorts (Netflix). SaaS Revival. 5-Stunden-Training-Switch | blog/2026-06-19_steve-yegge-flat-curve-society.md | | [AI Value Migration — Orchestration + Infrastructure](concepts/llm/ai-value-migration-orchestration.md) | 20VC mit Aravind Srinivas: Wert verschiebt sich von Modellen zu Orchestrierung, Strom, Mindset. Token Value per Watt per User als Schlüsselkennzahl. Yegge-Flat-Curve-Schicht ergänzt | xpost/2026-06-15_harrystebbings-20vc-aravind-srinivas.md + 2 | diff --git a/wiki/institutions/moonshot-ai.md b/wiki/institutions/moonshot-ai.md index 5a7c398..f46c7d6 100644 --- a/wiki/institutions/moonshot-ai.md +++ b/wiki/institutions/moonshot-ai.md @@ -1,7 +1,7 @@ --- created: 2026-06-24 -updated: 2026-07-19 -sources: [`raw/other/2026-06-13_kimi-k2.7-code-ollama.md`, `raw/youtube/2026-06-14_fahd-mirza-kimi-k2.7-vs-glm-5.2.md`, `raw/xpost/2026-07-18_healthranger-kimi-k3-anthropic-panic.md`, `raw/xpost/2026-07-19-bridgemindai-moonshot-capacity.md`] +updated: 2026-07-20 +sources: [`raw/other/2026-06-13_kimi-k2.7-code-ollama.md`, `raw/youtube/2026-06-14_fahd-mirza-kimi-k2.7-vs-glm-5.2.md`, `raw/xpost/2026-07-18_healthranger-kimi-k3-anthropic-panic.md`, `raw/xpost/2026-07-19-bridgemindai-moonshot-capacity.md`, `raw/xpost/2026-07-20-thebuggeddev-kimi-vs-fable.md`] tags: [institution, moonshot-ai, kimi-k3, capacity, go-to-market] --- @@ -37,6 +37,12 @@ Diese Entscheidung wurde als Markt-Signal wahrgommen: Customer-first vs. Growth- Im Juli 2026 veröffentlichte Moonshot AI **Kimi K3**, ein Frontier-Modell das ~8× günstiger ist als Claude-Äquivalente (~$15/M Tokens). Der Open-Source-Release ist für den 27. Juli 2026 angekündigt. Kimi K3 sorgte für Kontroversen wegen seiner ungefilterten Antworten im Vergleich zu Claude's Safety-Guardrails — ein strukturelles Dilemma zwischen US-Safety-First und Chinese-Utility-First. Siehe [[../concepts/llm/kimi-k3.md]] für Details. +### Real-World Benchmark: Vibe Engineering (20. Juli 2026) + +@thebuggeddev verglich Kimi K3 und Fable 5 in einer 3D-Fußballstadion-Coding-Challenge. Fable 5 war 3× schneller (under 1h vs. ~3h), aber Kimi K3 nutzte die Zeit für E2E-Tests, Responsive-Validierung (Desktop/Tablet/Mobile), Screenshots, Fehlererkennung und Low-End-Hardware-Anpassung. Kimi lieferte eine saubere React + Three.js-Architektur; Fable erzeugte eine monolithische HTML-Datei. + +**"Vibe Coding" (Fable 5) vs. "Vibe Engineering" (Kimi K3)** — das Extra an Zeit ist Qualitätsinvestition, nicht Langsamkeit. Dies verstärkt Moonshots Signal: Customer-first im Business Model + Engineering-Thoroughness im Produkt. Siehe [[../concepts/llm/kimi-k3.md]] für Details. + ## Cross-References - [[../concepts/llm/kimi-k3.md]] diff --git a/wiki/log.md b/wiki/log.md index 74601af..4b62431 100644 --- a/wiki/log.md +++ b/wiki/log.md @@ -2,6 +2,22 @@ *Append-only changelog. Start: 2026-06-05* +## [2026-07-20] Ingest | @thebuggeddev Kimi K3 vs. Fable 5 Real-World Coding Benchmark + +**Type:** ingest | **Scope:** raw/xpost, wiki/concepts/llm (3 updated), wiki/institutions (1 updated), wiki/index, wiki/log +**Source:** X-Post von @thebuggeddev — https://x.com/thebuggeddev/status/2079185808305823940 (20.07.2026, 46.5K views, 240 bookmarks) +**Trigger:** Subagent task: Wikify Kimi K3 vs. Fable 5 comparison post. +**Actions:** +- raw: `raw/xpost/2026-07-20-thebuggeddev-kimi-vs-fable.md` (created — 4.2 KB; Frontmatter [type: xpost, author: @thebuggeddev, is_thread: false, like_count: 464, bookmark_count: 240, tags: kimi-k3, fable-5, vibe-coding, vibe-engineering, coding-benchmark, real-world-test, react, threejs, e2e-testing, responsive, quality]. Content: Post summary, links, key takeaways, Hector analysis with combined Moonshot signals, 7 cross-refs) +- wiki: `concepts/llm/kimi-k3.md` (updated — added "Real-World Benchmark" section with comparison table, Vibe Coding vs. Vibe Engineering framework, combined Moonshot signals analysis, 3 new cross-refs) +- wiki: `concepts/llm/vibe-coding-vs-enterprise.md` (updated — added cross-ref to kimi-k3.md for Vibe Engineering distinction) +- wiki: `concepts/llm/fable-5-anthropic.md` (updated — added cross-ref to kimi-k3.md benchmark) +- wiki: `institutions/moonshot-ai.md` (updated — added Vibe Engineering benchmark section, new source in frontmatter) +- wiki: `index.md` (updated — Header auf "80. Update", Kimi K3 entry expanded with Vibe Engineering benchmark) +- log: this entry +**Hector-Hauptthese:** Der Benchmark bestätigt Moonshots Emerging Pattern: Thoroughness over Speed. Die Extra-Zeit ist Qualitätsinvestition — E2E-Tests, Responsive-Validierung, saubere Code-Architektur. Die "Vibe Engineering" vs. "Vibe Coding" Distinktion erweitert Tielke's Vibe-Coding-vs.-Enterprise-Taxonomie: der Split existiert nicht nur zwischen amateur und enterprise, sondern auch *innerhalb* KI-gestützter Entwicklung selbst. Kombiniert mit dem Capacity Event (19.07.) ergibt sich ein kohärentes Moonshot-Bild: Customer-first im Business Model + Engineering-Thoroughness im Produkt. +**Subagent-Modell:** ollama/glm-5.2:cloud + ## [2026-07-18] Ingest | @_The_Prophet__ AI Civilizational Integration **Type:** ingest | **Scope:** raw/xpost, wiki/concepts/agi (1 new), wiki/index, wiki/log