AN INDEPENDENT DATA STORY01—31 OCTOBER 2026
A month the internet
would rather forget.
An exposed population register. Critical flaws in PHP. A $6.36 billion crypto theft. Thirty-one days of cybersecurity, seen through Hacker News.
Explore the monthTHE MONTH IN THE RED01 — 31 OCT
0108152231
EVERY COLUMN, ANOTHER DAY.↗ EXPLORE
* HACKER NEWS. LIVING UP TO ITS NAME.SCROLL TO FOLLOW THE STORY
SECURITY STORIES
169worth a closer look
HN POINTS
45.8Kcollective attention
COMMENTS
24.5Kin the conversation
DAYS WITH A STORY
30/31observed on the front page
01 THE STORY SO FAR
Different headlines.
The same unease.
We trust strangers with our personal data, our servers, our money. This October brought reminders of what happens when that trust breaks. Three stories put the stakes in focus.
02 THE SHARE OF ATTENTION
Attention came
in waves.
Security stories first seen on October 21 account for 40.9% of their daily cohort’s recorded HN points. More than two in five. Was that a sustained drumbeat or a single spike? Trace the full month below.
OCT 21 / PEAK DAILY SHARE40.9%
of recorded HN points in this day’s group went to security stories
8,625 OF 21,103 POINTS
24 OF 107 STORIES
SECURITY’S SHARE OF RECORDED POINTS0–50%
01 OCT08152231
9.4%of the month’s recorded points
45,832 security points out of 488,935 across the full archive. The peak is one day, not the whole month.
Daily groups use the first observed front-page date. Points are the scores saved by the archive, not a count of votes cast that day. Inspect the complete calculation
03 CONNECT THE DOTS REAL STORIES. OPEN DATA.
A month under the microscope.
Find the patterns. Follow the stories. Make up your own mind.
SECURITY STORIES / DAYHOVER TO DISCOVER. CLICK TO EXPLORE.
WEEK 01WEEK 02WEEK 03WEEK 04OCTOBER
169 STORIES. EACH ONE A THREAD TO FOLLOW.
How to read thisBreachesVulnerabilitiesMalwareSupply chainPrivacyDefense
ATTENTION ≠ SEVERITY
The stories169
DATESTORY / SOURCEPOINTSCOMMENTS
169 matching stories
SHOWING 01—06 OF 169 STORIES
FROM THE HACKER NEWS ARCHIVE
Better at finding flaws.
Better at fixing them?
In April, Project Glasswing put Claude Mythos Preview in defenders’ hands to find and fix weaknesses in critical software. The premise was a head start for defense.
By October 8, Anthropic said it had not yet achieved enough reduction in cyber risk. Finding flaws was easier; verifying, prioritizing, and fixing them remained hard.
A patch announcement and a stolen wallet sit side by side in this archive. They are different outcomes of the same race: can defenders turn discoveries into deployed fixes faster than attackers turn them into exploits? These headlines cannot establish which side is winning, or that AI caused an attack.
FINDING A FLAW IS NOT THE SAME AS CLOSING IT.
05 THE SMALL PRINT
Good stories.
Transparent sources.
ARCHIVE SNAPSHOT
01 NOV 2026 · 05:15 UTC
What’s included? +
We read every available daily file in hckr news’s front-page archive, plus its live page for days not archived yet. For October 2026, that is 2,898 unique story records across all 31 days. There is no minimum score. Dates are first observed front-page appearances in UTC, not submission dates. Jobs are excluded. Different stories can cover the same incident.
Does this capture every brief appearance? +
It captures all matching records returned by the archive, including low-scoring stories and those submitted before October. That is broader than the original daily top-30 sample. It still cannot prove that a story appearing only between polls was captured, or that the archive had no outages. Where available, each story also links to an independent HN Rankings record with actual top-30 observations. Missing rank history remains missing; no time on the front page is inferred.
How is the attention share calculated? +
For each UTC day, we sum the source’s saved HN scores for security-related stories and divide by the saved scores for every recorded non-job story first seen that day. The monthly figure uses the sum across the full month, not an average of daily percentages. HN points are a proxy for attention, not raw upvote counts, unique voters, or votes cast that day. Scores are snapshots. Inspect every record in the denominator.
How are stories classified? +
Title and URL keywords identify six themes, including defensive work and digital privacy alongside attacks. Explicit, documented corrections exclude unrelated meanings such as job security, hardware hobby projects, or a contractual breach. Each story has one primary category. Security-related here includes research and policy, not only confirmed attacks. Classification can still miss context; the original sources are one click away.
What do the numbers mean? +
Points and comments show attention, not damage. They are the counts saved by the source; currently featured stories can be updated on refresh. The attention analysis stays fixed to its saved archive snapshot. 24 stories on October 21 is the busiest recorded daily count for 2026. Some headlines concern historical incidents, research, or policy rather than an attack that happened in October.