Feeding a Model 600,000 Real Websites to Roast Its Own Landing Page—Then Rewriting It
PrajwalTomar_ · x · 2026-09-18
Developer Prajwal Tomar connected his model to an MCP server with 600,000+ real websites and had it roast the landing page it had just built for him. The verdict: the page looked expensive but had zero substance—stacked decorations and interchangeable copy, which he calls the most accurate roast of AI-generated websites he's read. After saying "approved," the model rewrote the entire page with visibly better results. His point: most people type "make it look premium" into Cursor or Claude Code and wonder why everything looks like AI slop. Full workflow, prompts, and the report are in his article.
Related event: GPT-6 Astra redesigns landing pages using 600k real websites(2 posts)→
More from coding & agent
- DeepSeek design doc: a person can only follow 3-5 agent sessions at once, attention is the real bottleneck — heyshrutimishra · 2026-11-02
- Software fundamentals are key to getting good results from AI coding agents — bendee983 · 2026-11-02
- Open-source LLM regression gate digline hits 1.0: quantifying judge noise before calling a regression — alexpran · 2026-11-02
- PR adds patient retry logic for model capacity errors in OpenAI Codex — bytebot · 2026-11-02
- Perplexity open-sources WANDR: 500 research tasks, 170,495 evidence points, URL-rechecking grader — thisguyknowsai · 2026-11-02
- 4 AIs, 96 questions, 2,671 citations: 74% are blog posts, 99% on companies' own sites — tibo_maker · 2026-11-02