I'm building an AI-proof writing tool. How would you defeat it?
Details
- External ID
- 46799402
- Source
- HN
- Company
- —
- Product
- I'm building an AI-proof writing tool. How would you defeat it?
- Website domain
- vercel.app
- Launched
- Jan. 28, 2026
- Cohort
- —
- Upvotes
- 24
- Upvotes percentile
- 0.6811594202898551
- Tags
- —
- Fetched at
- Sept. 7, 2026, 9:25 p.m.
- Updated at
- Sept. 7, 2026, 9:25 p.m.
Enrichment
- Theme
- AI text humanizers and detectors
- Vertical
- Horizontal
- Function
- Content generation
- Audience
- Developer
- AI stance
- AI feature
- Project type
- Commercial product
- Normalized one-liner
- ai-detection resistant writing tool
- Manually corrected
- False
Could you build this?
Partial Building a browser text editor that collects keystroke dynamics and tab blur events is easy, but reliably distinguishing genuine human typing from simulated or transcribed AI text is an unsolved adversarial problem.
What it would actually take: The frontend captures flight time, dwell time, inter-key latency, jitter, and DOM events, streaming them to a backend running statistical anomaly detection and classification models trained on human vs. synthetic behavioral biometrics. The difficult challenge is preventing replay attacks via synthetic hardware injection (e.g. Arduino USB HID or low-level CDP input events) and avoiding high false-positive rates for non-native or atypical typists. Requires machine learning and security/biometric anti-cheat expertise.
Discussion
20 comments analyzed.
Competitors mentioned: Gemini, ChatGPT
Concerns raised: Easy to defeat by copy-typing LLM output with manual keyboard input, Impossible to prevent cheating if user is motivated; hardware/software keystroke emulation available, Statistical approach can be spoofed with recorded keystrokes and pauses replayed via xdotool, Eye tracking can be spoofed with masks or webcam spoofing; accessibility issues with eye tracking requirement, Window/tab switching detection irrelevant when windows placed side-by-side
Feature requests: Detect backspaces and cursor movement patterns in addition to keystroke timing, Implement eye tracking to verify gaze matches screen activity, Increase penalties for tab/window switches during composition, Require webcam verification with eye tracking
Competitors
Other products that read as similar to this one — 563 launches clear the similarity bar, closest 8 shown.
Attention rank: #156 of 564 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 91 days after the earliest competitor.
- Can you really detect AI writing from human writing? · hn · 2026-05-13 · 7 upvotes · similarity 0.60
- snifftest · github · 2026-09-17 · 27 upvotes · similarity 0.56
- writ · github · 2026-09-21 · 9 upvotes · similarity 0.54
- AI writes code · hn · 2026-02-21 · 8 upvotes · similarity 0.54
- learning_ai_pentesting · github · 2026-09-26 · 24 upvotes · similarity 0.54
- Academic-Writing-Humanizer · github · 2026-09-26 · 7 upvotes · similarity 0.53
- Xalgorix · hn · 2026-08-19 · 9 upvotes · similarity 0.52
- Kasspian · hn · 2026-06-11 · 5 upvotes · similarity 0.52
Other launches for this product
- No other launches for this product.
Same idea, different domain
Nobody's really built a content generation tool for Government yet.