AI infrastructure on Hacker News
What Hacker News has posted about AI, GPUs and data centers over the last 30 days: how the topics compare, which get the most attention, and who and what the stories are about. Updated every day.
Key figures
Stories and counts come from Hacker News through its public search API. As of 2026-10-26T01:11:53+00:00.
- All topics: 1,013 stories this week, 1,061 the week before (−5%), as of 2026-10-26T01:11:53+00:00.
- Busiest hour for all topics: Tuesdays around 14:00 UTC, about 15.8 stories an hour, over the last 30 days. Hours are in UTC.
- Google: 17 titles this week, 16 the week before.
- Every AI model has the same "controversial" take on humans (modelcompare.ai): 655 points, 634 comments.
Rest of the figures
What Hacker News has posted about AI, GPUs and data centers over the last 30 days: how the topics compare, which get the most attention, and who and what the stories are about. Updated every day. Stories and counts come from Hacker News through its public search API and are refreshed every day. A story counts toward a topic when the topic appears in its title, so one story can count toward several.
Topics, stories posted.
All topics: 1,013 stories this week, 1,061 the week before (−5%). Over the last 30 days, 4,384 stories, 52,430 points and 35,155 comments, as of 2026-10-26T01:11:53+00:00.
AI: 940 stories this week, 982 the week before (−4%). Over the last 30 days, 4,089 stories, 46,460 points and 31,427 comments, as of 2026-10-26T01:11:53+00:00.
AI agents: 202 stories this week, 173 the week before (+17%). Over the last 30 days, 779 stories, 7,166 points and 3,559 comments, as of 2026-10-26T01:11:53+00:00.
GPUs: 50 stories this week, 45 the week before (+11%). Over the last 30 days, 172 stories, 2,604 points and 739 comments, as of 2026-10-26T01:11:53+00:00.
Data centers: 45 stories this week, 54 the week before (−17%). Over the last 30 days, 187 stories, 4,955 points and 4,190 comments, as of 2026-10-26T01:11:53+00:00.
This week, all topics: 1,013 stories, 1,061 the week before.
Companies named in titles this week. A company counts when its name or a main product (ChatGPT, Claude, Gemini, Copilot, Llama, Grok, AWS) is in the title.
Nvidia: 16 titles this week, 18 the week before.
Amazon: 11 titles this week, 5 the week before.
Meta: 10 titles this week, 12 the week before.
Anthropic: 9 titles this week, 27 the week before.
OpenAI: 7 titles this week, 13 the week before.
DeepSeek: 5 titles this week, 0 the week before.
Microsoft: 5 titles this week, 6 the week before.
xAI: 3 titles this week, 0 the week before.
Samsung: 2 titles this week, 0 the week before.
Top stories this week, most upvoted first.
Ask HN: How to deal with shipping AI-written code you don't understand? (Hacker News): 620 points, 762 comments.
No signs of AI payoff outside the tech sector (apollo.com): 610 points, 645 comments.
America has lost 200k of jobs in media, film and the arts. How much is AI? (apricitas.io): 503 points, 542 comments.
OpenShell – A safe, private runtime for fleets of autonomous AI agents (github.com): 501 points, 218 comments.
PHP2Shell: AI uncovers critical RCE in PHP (wiz.io): 454 points, 185 comments.
Cerebras Deploys Trillion-Parameter AI Models (cerebras.ai): 352 points, 127 comments.
GPU-unlock: FP/BF16 instruction issue rate on Nvidia 3000's cards (github.com): 337 points, 80 comments.
A practical guide to AI image generation, from diffusion to agents (danielball.com): 283 points, 94 comments.
Google illegally cut down 250K trees in Finland for data center (nypost.com): 258 points, 209 comments.
Why is AI generated language so hard to understand? (arxiv.org): 241 points, 197 comments.
AI use leads UC Berkeley math professor to see 'no point' in grading homework (dailycal.org): 199 points, 284 comments.
Steam Users Beg for "No AI" Filter as Vibecoders Flood Next Fest (ign.com): 184 points, 231 comments.
Using thousands of AI agents to uncover new leads in biological data (openai.com): 160 points, 106 comments.
Show HN: Qwen3-4B with the AI writing tics trained out (100toolkit.com): 158 points, 113 comments.
The same AI request emits 8.5 times more CO2 in Texas than in Sweden (berget.ai): 142 points, 157 comments.
AI and the Enclosure of the Scientific Commons (arxiv.org): 141 points, 81 comments.
Nikon photography contest winner loses title after scientists spot signs of AI (techspot.com): 132 points, 71 comments.
Google staff resign in backlash over its AI contracts with military (bbc.com): 128 points, 26 comments.
AI-Assisted Paper on Z-Library Cites a TorrentFreak Article That Doesn't Exist (torrentfreak.com): 97 points, 61 comments.
On AI writing (bolkunov.com): 96 points, 62 comments.
Imagine an AI future where being banned means you are now economically useless (twitter.com): 94 points, 74 comments.
TornadoVM 7 Enables GPU-Native Workloads from Java Applications (infoq.com): 94 points, 22 comments.
Most Americans Say AI Is Bad for Society, Oppose More Data Centers (abcnews.com): 91 points, 68 comments.
AI analysis of CT scans finds 42.9% of patients have coronary inflammation (biograph.com): 91 points, 47 comments.
Show HN: UI micro interactions prompts for AI (typeui.sh): 80 points, 46 comments.
Rust off Rails: The opinionated web stack for the AI era (rustoffrails.org): 76 points, 76 comments.
LayerFS: A filesystem for AI coding agents (layerfs.ai): 76 points, 47 comments.
Show HN: We had AI write a wiki from 45k Enron emails, every fact cited (thisandthat.chat): 70 points, 56 comments.
Award-winning AI literature (jamesyu.org): 66 points, 76 comments.
Topics
Pick a topic to see its companies, sources and stories further down.
Each row's bars are scaled to that topic's own busiest day. The chart below puts the topics on one scale.
Topics compared over time
AI is left out at first because it is many times larger than the others. Select it in the legend to add it.
Attention per story
Average over the last 30 days. A topic can be small and still draw more votes per story.
Busiest hours of the week
This week
Companies named in titles
A company counts when its name or a main product (ChatGPT, Claude, Gemini, Copilot, Llama, Grok, AWS) is in the title.
Where the stories come from
Top stories
Loading stories…
In the press
Stories and counts come from Hacker News through its public search API and are refreshed every day. A story counts toward a topic when the topic appears in its title, so one story can count toward several.