Before LLMs
Build a site. Optimize for SEO. Maybe run some ads. You’d get a trickle of visitors, some traction. Whether it was a static blog or a dynamic app, you had a shot at attention, or even income.
Enter LLMs
Now? Any public content you publish online is fair game for large language models (LLMs). They read it, regurgitate it, and serve it in sleek interfaces. Sure, their training data may lag a year or two, but they’re getting better at browsing and parsing content in real time.
Why People Prefer LLMs
Let’s be honest, most websites are bloated messes. Bad typography, clunky navigation, popups, buried content. LLMs cut through all of it. Ask, and they deliver a clean, summarized answer. Why dig through 10 tabs when an AI assistant can scan the web and hand you the answer?
It’s not always accurate, and it doesn’t always cite sources well, but convenience is winning. Your visitors are disappearing, and the bots aren’t paying you for your knowledge.
So, Should You Even Bother?
Yes, but only if you make it worth visiting. Try this instead:
- Build interactive content: Let users engage, play, or learn something by doing.
- Design something memorable: Think weird, funny, beautiful, just different.
- Go app-like: If necessary, hide content behind auth. But do this sparingly, it's always better to provide value immediately.
What's Next?
Someday, LLMs might render and serve websites inside their answers, iFrame-style. They’ll remix human-made ideas into something interactive and custom. Until then, your job is to be the source worth linking to, not just the one feeding the machine.