Going semi-viral on Hacker News

My article went semi-viral (if you could call it that) on Hacker News in January 2026.

At one point, it reached the first page of the home page, rising up to rank #27, which felt huge for me as a casual HN reader and new blogger. Apparently the site got 13+ million visits in that same month¹.

I've never had my writing receive this much attention before, so I wanted to share what that experience was like.

Reception

The post received 280+ comments on HN, and the general sentiment was mixed. Some agreed with my thesis. Some didn't. Some were combative.

Initially, the negative comments felt personal and surprisingly unnerving. This gave me a newfound respect for bloggers like Sean Goedecke, who publish frequent, popular opinion pieces.

Outside of HN, the post also got shared on other platforms by many people, notably by Google Cloud's Chief Evangelist, a Computer Science professor, and a popular tech Youtuber. It felt rewarding to see my ideas resonating with others.

Analytics

During that first week, this post saw 13k unique visitors, 14k sessions, and 16k pageviews, with a bounce rate of 92% and visit duration of 34 seconds.²

Visitors, sessions, and pageviews being close to a 1:1:1 ratio and high bounce rate meant most people came to read that post then immediately left. The 34-second visit duration meant an average visitor read 47%, or almost half, of the post.³

These behaviors match my own experience when browsing HN: I'd click on a post and read the first few sentences. If the post doesn't immediately capture my interest or seems too long, I'd leave right away.

Costs

Going semi-viral on HN almost took down this website.

I'm using Vercel to host this blog under their Hobby (free) plan. Soon after the post got popular, Vercel emailed me that I have "used 100% of the included free tier usage for Edge Requests (1,000,000 Requests)."

The number of requests was much higher than pageviews because of the assets (images, fonts, etc.) that needed to be loaded every time. Luckily, I put in some caching strategies that reduced the request:pageview ratio.

I'm still a (skeptical) Vercel fan, but this made me weary of using them for hosting any serious projects in the future.

Footnotes

1. Hacker News' estimated traffic data came from SimilarWeb. ↩

2. I use Umami for very basic website analytics. ↩

3. Assuming an average reading speed of 238 wpm, the average reader read 34 seconds * (238 words / 60 seconds) = 135 words, which is ~47% of the post's 287 total words. It's also likely that HN readers have a much faster wpm, and they only skimmed the article instead of reading the post in its entirety. ↩