AI just profiled someone from 1,000 public comments — in seconds

Simon Willison — one of the most respected voices in AI development and co-creator of Django — just demonstrated something that should make anyone with a public internet presence uncomfortable. He fed 1,000 of his own Hacker News comments into Claude and asked it to "profile this user."

The result? A disturbingly accurate portrait that identified his name, location, career history, technical opinions, personality traits, and even his debate style. He called it "startlingly effective" — and "mildly dystopian."

How It Works — Three Steps, Zero Coding

The technique is alarmingly simple. Willison built a free web tool that anyone can use right now:

That's it. No coding, no special tools, no hacking. Everything uses publicly available data that Hacker News users voluntarily posted.

What the AI Found

When Willison ran the technique on himself, Claude correctly identified:

Claude likely identified Willison by name because he links to his personal website in comments. But even without a name, the profile would have been remarkably specific — enough to narrow down a person's identity, employer, and beliefs.

Why This Should Worry You

This technique works on any platform where comments are public — not just Hacker News. Reddit, Twitter, GitHub discussions, forum posts, product reviews. Anywhere you've left a trail of text, AI can now stitch it into a profile.

What makes this different from traditional "doxxing" (searching for someone's personal info online):

• Speed: What would take a human hours of reading takes AI about 30 seconds
• Pattern recognition: AI spots personality patterns, political leanings, and lifestyle details that humans would miss across hundreds of posts
• Scale: Anyone can do this, for free, right now — no technical skills needed
• Aggregation: AI combines tiny clues across many comments into insights no single comment would reveal

Willison himself acknowledged the technique is "invasive" — even though it only uses data people chose to make public.

Try It Yourself

If you want to see what AI can learn about you (or understand what others could learn), here's how:

# The API endpoint that fetches any user's comments:
https://hn.algolia.com/api/v1/search_by_date?tags=comment,author_USERNAME&hitsPerPage=1000

# Or just use Willison's tool:
https://tools.simonwillison.net/hn-comments-for-user

Paste the results into Claude, ChatGPT, or any AI chatbot and ask it to "profile this user." The results may surprise you.

The Bigger Picture

This isn't just a parlor trick. It reveals something fundamental about how AI changes the value of public data. Comments that felt anonymous and scattered across years of posting can now be instantly aggregated and analyzed.

For Willison, the practical use case is mundane: checking whether someone arguing with him on Hacker News has a history of bad-faith comments before engaging further. But the same technique could be used by employers screening candidates, advertisers building profiles, or anyone curious about someone's digital footprint.

The uncomfortable truth: the data was always public. AI just made it readable at scale.

Related Content — Get Started with Easy Claude Code | Free Learning Guides | More AI News