18.7% of Popular Hacker News Stories in 2026 Are About AI, Up From 1.9% in 2021

Summary

Picture a software founder who opens Hacker News every morning to see what the tech crowd cares about. In 2021 they would scroll past maybe one AI story a day. Today it is closer to one in every five popular links.

Hacker News is the discussion site run by startup accelerator Y Combinator. Engineers, founders and product people vote stories up, and the ones that pass about 100 points usually spend time on the front page. That makes it a useful early read on where technical attention is going before the rest of the market follows.

We labeled every story that reached 100 points from 2015 through September 2026 as AI or not, by matching its title against a list of AI and machine learning terms. In 2026 so far, 18.7% of them were. That is nearly ten times the 2021 share, and the trend line has not flattened.

What we measured

We used the public Hacker News table on Google BigQuery, which holds every story and comment. We kept live stories posted from January 2015 to September 2026 that scored at least 100 points. The table was refreshed through 17 October 2026, so September 2026 is the last full month.

Each title was labeled AI if it matched a list of AI and machine learning terms, such as AI, LLM, GPT, machine learning, neural, agents and the names of major model families. Our team checked a random sample of 200 titles labeled AI and 200 recent titles labeled not AI. About 2 of 200 AI labels were wrong (for example a programming language called ML), and about 7 of 200 non-AI titles were AI-related stories the list missed. We added terms for the missed cases, but the real share is likely a little higher than reported.

AI went from 2% to nearly 19% of popular stories in five years

From 2015 to 2022 the AI share of popular stories sat between 1.4% and 3.3%. Machine learning research and deep learning tools showed up, but they were a niche. Then 2023 arrived.

The share jumped from 3.0% in 2022 to 10.3% in 2023, held near 10% in 2024, then climbed to 14.4% in 2025 and 18.7% in the first nine months of 2026. The 2026 figure covers January to September only.

The total number of popular stories fell from 17,767 in 2023 to about 14,000 a year since. AI stories dipped in 2024 too, from 1,834 to 1,363, then grew in count as well as share in 2025 and 2026.

The biggest months are the most recent ones

Monthly data shows two waves. The first followed the ChatGPT launch at the end of November 2022. The second started in spring 2025 and is still building.

The first wave peaked in March 2023 at 15.1% and then cooled to about 7% by September 2023. The second wave has not cooled. Every month from February to September 2026 came in above 18%.

AI stories score higher and pull more comments

Getting to 100 points is only the first step. Once there, AI stories in 2025 and 2026 kept climbing higher than other stories and started longer arguments.

Before 2022, AI stories trailed other popular stories on both points and comments. Points flipped in 2022 and comments in 2024, and both gaps have widened since. In 2026 popular AI stories averaged 21.0% more points and 34.2% more comments.

The top of the board shows the same pattern. Of the 1,785 stories that passed 500 points in 2026, 460 (25.8%) were about AI.

What this means

Hacker News readers are the people who build and buy software, and about a fifth of the stories they push to the front page are now about AI. The same shift is reaching your customers a year or two later, in how they search, compare and pick vendors. When they ask an AI assistant for a recommendation, the answer depends on what that assistant has read about you. Shaping what assistants read about you is the job of an AI search optimization agency, and it starts with your own site.

You do not need to post on Hacker News to act on this. Check how AI tools describe your business today and fix what is missing or wrong while the habit is still forming. Our free AI visibility audit is an easy place to start; think of it as the check an AEO agency runs on day one, done before you commit to anything.

Limits

A 100-point score is a stand-in for front-page placement. Hacker News ranking also depends on time, flags and moderation, which the dataset does not record.

AI labels come from title keywords checked on a random sample of 400 titles. Roughly 1% of AI labels were wrong and about 3.5% of non-AI titles were AI-related, so the true AI share is probably slightly higher than reported.

The AI label is broad and includes machine learning research, AI products, AI policy and AI coding tools.

Scores and comment counts are the values at the time of the October 2026 refresh. 2026 covers January to September only.

Hacker News readers skew technical, so the trend shows attention among developers and founders, not the general public.

Frequently asked questions

Related studies

• Developer conversation on Hacker NewsAI SEO Draws 3x the Positive Developer Comments of SEO
• Developer conversation on Hacker Newsllms.txt Splits Developers: 54% For, 46% Against
• Developer conversation on Hacker News'Written by AI' Comments on Hacker News Up 4.6x Since 2023
• Developer conversation on Hacker News'Slop' Now Shows Up in 1 of Every 149 Hacker News Comments