I tracked the AI obsession on Hacker News and the numbers are

I spent some time digging through Hacker News archives to see just how much the platform has been swallowed by the current LLM hype cycle. It isn't just a feeling; the data actually shows a massive, accelerating surge in how often these topics appear in headlines. I built a little tool to quantify this, specifically looking for the standalone word "AI" to keep the data clean, though I also ran an extended version to capture the broader ecosystem.

The methodology is pretty specific. If a title says "AI", it counts. If it says "OpenAI", it doesn't (unless you toggle the wider search). This distinction is important because it separates the generic buzzword from specific brand mentions. When you toggle the extended vocabulary, it pulls in "artificial intelligence", "LLM", "GPT", and specific model or vendor names.

Here is what the data actually looks like right now:

• Current Standard AI Density: 15% of today's titles contain the word "AI".
• Current Extended Density: 19% when including LLMs, GPT, and specific vendors.
• Year-over-Year Trend: 14.4% of new HN titles this year compared to 10.9% in 2025.
• Extended Vocabulary Trend: 21.9% this year versus 16.1% in 2025.
• Category Peak: The "Show HN" category is currently leading the pack at 19.8% AI mentions.

Looking at the historical timeline, the growth isn't linear; it's explosive. We didn't even see a single month cross the 1% threshold until October 2016. It took almost seven years to hit 5% (which happened in February 2023), and we only just crossed the 10% mark in May 2025.

If you want to look at the volatility, the five-year range is massive. The absolute floor was January 2022, where AI mentions sat at a measly 0.8%. On the flip side, the peak hit 15.9% back in February 2026. It's also interesting to note the weekday vs. weekend split: AI topics make up about 13.6% of weekday titles compared to 12.4% on weekends, suggesting that the "work" side of the AI hype is slightly more concentrated during the business week.

It's a bit of a meta moment because this specific post, being a "Show HN" submission, actually contributes to the very statistics I'm reporting. It really highlights how much the discourse has shifted from general software engineering and distributed systems toward prompt engineering and model deployment. Whether this is a permanent shift in the tech landscape or just a massive bubble that will eventually deflate is still anyone's guess, but for now, the signal is undeniably loud.