State of Developer Conversation 2026

Bluesky out-talks Hacker News

We tracked public conversation about 48 developer tools for 30 days. With bots and feeds removed, Bluesky still carried 1.9x the conversation of Hacker News. Stack Overflow got 968 new questions in October.

Octolens research · Published · Data from Sep 30 to Oct 29, 2026

Bluesky against Hacker News
1.9x

Bluesky against Hacker News

Relevant dev-tool mentions a day, with bots and feeds removed.

New Stack Overflow questions
968

New Stack Overflow questions

In October 2026. October 2015 had 186,031.

Non-AI tool mentions that bring up AI
1 in 3

Non-AI tool mentions that bring up AI

We checked this against a labeled sample of 600 mentions.

Relevant mentions a day, 48 developer tools

PeopleBots and feeds
GitHub issues1,337 a day
X1,330 a day
Reddit893 a day
Bluesky430 people · 208 bots and feeds
News235 a day
Hacker News227 a day
TikTok96 a day
DEV35 a day
Sep 30 to Oct 29, 2026. We measured automated accounts on Bluesky and Hacker News only. Volumes are minimums for the largest platforms. Podcasts, YouTube and LinkedIn left out because coverage was partial.

Finding 1 · The rooms

Bluesky out-talks Hacker News

1.9x

as many relevant dev-tool mentions a day on Bluesky as on Hacker News, bots and feeds removed

Across 48 developer tools, Bluesky logged 638 relevant mentions a day and Hacker News logged 227. Before cleanup, that looks like nearly 3x. Status-page relays, job boards, trending-repo feeds and accounts that repost Hacker News made up 33% of Bluesky's volume, so we removed them. People alone still posted 430 relevant mentions a day.

The two rooms talk differently. 82% of Hacker News mentions are comments under a story, so the talk clusters around the links that reach the front page. On Bluesky, 74% of what people post are standalone posts, not replies.

We counted posts, not readers. One Hacker News front-page thread can reach more people than hundreds of Bluesky posts, so this measures how much developers talk, and says nothing about influence. Depending on how strictly we define a bot, the ratio lands from 1.9x to 2.1x.

relevant Bluesky mentions a day from people
430
relevant Hacker News mentions a day
227
of Bluesky's volume came from bots and feeds
33%

“It’s been absolutely delightful building a Svelte-assed Svelte app with SQLite and Litestream backups.”

Post on Bluesky, October 2026

How each room talks, relevant mentions a day

PostsReplies
Bluesky318 posts · 112 replies
Hacker News227 a day
Bluesky with bots, feeds and Hacker News mirror accounts removed. On Hacker News, under 1% of mentions came from automated accounts.

With bots removed, Bluesky carried 1.9x the developer-tool conversation of Hacker News in October 2026.

State of Developer Conversation 2026

Finding 2 · The decline

Stack Overflow went quiet

968

new public questions on Stack Overflow in October 2026

Is Stack Overflow dead? As a place to ask new questions, close to it. Developers asked 186,031 questions there in October 2015 and 968 in October 2026. The best month on record was March 2014, with 207,169.

The slide began years before AI chat tools. October counts fell 43% between 2015 and 2022. The fall sped up after ChatGPT launched in November 2022: down 51% in 2023, 56% in 2024, 77% in 2025 and 82% in 2026. We can show the timing, and this data can't prove the cause.

Our month of tracking agrees. Across 48 tools we logged about 9 Stack Overflow keyword matches a day, out of 15,408 across every platform. The archive stays online, and we didn't measure how many people still read it.

fewer questions than October 2015
99.5%
answers in October 2026, down from 137,375 in October 2022
2,310
of our keyword matches came from Stack Overflow
1 in 1,758

New questions on Stack Overflow each October

October 2026

968 questions

View the data
OctoberNew questions
2012149,907
2013183,850
2014172,631
2015186,031
2016175,229
2017169,634
2018160,372
2019152,587
2020141,749
2021118,763
2022105,743
202352,333
202423,161
20255,262
2026968
Public questions by creation date, deleted questions excluded. Source: Stack Exchange API, pulled Nov 9, 2026.

Stack Overflow received 968 new public questions in October 2026, down 99.5% from 186,031 in October 2015.

State of Developer Conversation 2026

Finding 3 · The questions

Where the questions went

46%

of the public product questions we found sit in GitHub issues

Developers still ask about their tools in public. We tagged 21% of relevant mentions as a product question, and almost none of them were on Stack Overflow.

A GitHub issue goes to the maintainers. A post on X goes to whoever follows you. Neither builds the shared archive of answers that Stack Overflow did.

We can't see the questions people ask ChatGPT or Claude. Those stay private, so this chart only covers questions asked in public.

of relevant mentions ask a product question
21%
of those questions are on X
20%
of those questions are on Reddit
19%

“What are you using for observability and monitoring in your AI projects? Datadog, LangSmith, Logfire, or something else?”

Post on X, October 2026

Share of public product questions

GitHub issues46%
X20%
Reddit19%
Bluesky6%
Hacker News3%
Podcasts2%
Estimated from each platform's volume and question rate. Volumes are minimums on the largest platforms.

46% of public product questions about 48 developer tools sit in GitHub issues, with 20% on X and 19% on Reddit.

State of Developer Conversation 2026

Finding 4 · The map

Where developers talk

72%

of relevant conversation happens on GitHub issues, X and Reddit

Three platforms hold most of the conversation about developer tools. GitHub issues and X are nearly level, and Reddit follows. Collection hit its limit on the biggest tools, so treat these shares as estimates.

Bluesky comes fourth. Even with its bots removed, it carries more than Hacker News, TikTok, DEV and LinkedIn combined (362 a day). We only measured automation on Bluesky and Hacker News, so this chart shows every platform before bot removal.

GitHub issues
27%
X
27%
Reddit
18%

“how is fly.io/ compared to something like cloudflare or digitalocean?”

Reply on X, October 2026

Share of relevant conversation

GitHub issues27.0%
X26.8%
Reddit18.0%
Bluesky12.9%
News4.7%
Hacker News4.6%
TikTok1.9%
DEV0.7%
Shares include 3.3% for podcasts, which we collected in part. Stack Overflow rounds to 0%. YouTube and LinkedIn could not be measured.

GitHub issues, X and Reddit carry about 72% of public conversation about developer tools.

State of Developer Conversation 2026

Finding 5 · The mood

Hacker News is the toughest room

19%

of non-bug Hacker News mentions are negative, the highest we measured

We removed bug reports before comparing platforms, because a bug report reads as negative by nature. Hacker News comes out as the harshest place to get discussed, and Reddit is about as positive as X (41% against 42%).

GitHub issues show why the adjustment matters. With bug reports included, 26% of GitHub mentions are negative. Without them, 7% are, and 83% are neutral working conversation.

TikTok is the friendliest platform, at 55% positive. It stays at 51% after we remove promotional posts.

negative on X and on Reddit
13%
positive on TikTok, the friendliest platform
55%
of GitHub bug reports read as negative
93%

“How much of the four months went to keeping up with breaking changes? Thats been my main blocker with zig, more than the language itself.”

Comment on Hacker News, October 2026

Sentiment by platform, bug reports removed

PositiveNeutralNegative
Hacker News28% positive · 19% negative
Bluesky31% positive · 15% negative
Reddit41% positive · 13% negative
X42% positive · 13% negative
Podcasts32% positive · 12% negative
TikTok55% positive · 9% negative
GitHub issues10% positive · 7% negative
News17% positive · 4% negative
Sorted by negative share. Podcasts are from partial collection.

Hacker News has the most negative conversation about developer tools: 19% of non-bug mentions, against 13% on X and Reddit.

State of Developer Conversation 2026

Finding 6 · The categories

Frameworks get the warmest talk

+16

net sentiment for frameworks and databases, the highest of eight categories

Developers talk about frameworks and databases more warmly than anything else we tracked. Observability and analytics tools sit close to even, at +2.

Hosting draws the most complaints, with 24% of mentions negative. We publish these numbers by category and never rank single tools.

net sentiment for observability and analytics
+2
negative for hosting, the highest of any category
24%
net sentiment for AI coding tools
+13

Net sentiment by category

Frameworks+16
Databases+16
AI coding+13
Team tools+12
Languages and runtimes+8
Hosting and infrastructure+5
Auth and payments+4
Observability and analytics+2
Positive share minus negative share, in points. Includes bug reports.

Frameworks and databases get the warmest conversation among developer tools. Observability gets the coolest.

State of Developer Conversation 2026

Finding 7 · The noise

Two in three matches miss

32%

of raw keyword matches were about the tool people searched for

We logged 15,408 keyword matches a day across the 48 tools. 4,959 of them were about the tool.

The name decides most of it. Clerk, Linear, Neon and Render all land under 10%, because people use those words for other things every day. Claude Code reaches 74% and Datadog 70%. No invented name fell below 20%.

Octolens AI made these relevance calls, so the rates reflect how our classifier reads each post.

median for names that are also ordinary words
15%
median for invented names
46%
ordinary-word names are below 20% relevance
15 of 24

Keyword relevance by tool

Hover a dot to see the tool

Ordinary-word namesMedian 15%
Invented namesMedian 46%
Each dot is one tool. Grouping is ours, and a few names are loose fits.

Only 32% of raw keyword matches for 48 developer tools were about the tool. For names that are also ordinary words, the median is 15%.

State of Developer Conversation 2026

Finding 8 · The AI share

AI is in a third of it

1 in 3

mentions of non-AI developer tools also bring up AI

We looked at the 42 tools outside AI coding: databases, payments, hosting, observability and the rest. About a third of their mentions also bring up AI, agents or a model by name. Passing references count.

An LLM labeled a random sample of 600 of these mentions, and we reviewed the hard cases by hand. Posts that bring up AI read warmer, at 32% positive against 21%.

Six of the 48 tools are AI coding tools. Together they take 22% of all relevant conversation.

share, with a range of 33% to 39%
36%
of all conversation goes to six AI coding tools
22%
of AI-related posts are positive, against 21% of the rest
32%

“I explained the situation to Claude and let it rip against Grafana logs.”

Post on X, October 2026

Sentiment, non-AI tool mentions

PositiveNeutralNegative
Brings up AI32% positive · 16% negative
No AI21% positive · 21% negative
Weighted sample of 600 mentions.

About a third of mentions of non-AI developer tools also bring up AI.

State of Developer Conversation 2026

Finding 9 · The pairs

The tools people compare

4.7%

of posts name two of the 48 tools

Claude Code and Cursor lead the list by a clear margin. Most other top pairs are stacks people build on together: Next.js with Vercel, Supabase with Vercel, Next.js with Supabase.

Direct rivals appear lower down, such as Datadog with Grafana and Lovable with Replit.

Posts naming both tools

Claude Code + Cursor186
Supabase + Vercel127
Next.js + Vercel126
Next.js + Supabase113
Cloudflare + Vercel101
Datadog + Grafana83
Lovable + Replit75
Netlify + Vercel71
Notion + Slack69
Rust + TypeScript66
From a sampled pool. Use for relative order, and don't read the counts as totals.

Claude Code and Cursor are the most co-mentioned pair among 48 developer tools.

State of Developer Conversation 2026

Also in the data

Smaller findings

  • 90%of the conversation is in English. Japanese comes second at 2.1%, ahead of Spanish at 1.7%.
  • 59%of public bug reports land in GitHub issues, more than every other platform combined.
  • 42%of Reddit mentions carry product feedback, the highest rate of any platform.
  • 32%of replies are positive, against 27% of top-level posts.

The panel

All 48 tools

Mentions per day are minimums for the largest tools, so read the order as a rough guide.

Tool
Claude Code38674%
Datadog15070%
GitHub Copilot18468%
Replit14466%
Cloudflare32266%
Grafana15962%
PostHog13462%
Auth05259%
PlanetScale3154%
Supabase18253%
Figma19152%
Windsurf7050%
Vercel17447%
Next.js15245%
Raycast8744%
TypeScript19643%
Rust25143%
Node.js17542%
WorkOS2641%
Svelte8639%
MongoDB11039%
New Relic2039%
Cursor19538%
Deno6938%
Netlify10237%
Nuxt5135%
Fly.io3332%
Slack14030%
Stripe13230%
Lovable9828%
Sentry8827%
Zig8822%
Notion10020%
Postman5818%
Angular7418%
Railway7416%
Prisma4915%
Convex4515%
Astro3510%
Axiom288%
Render428%
Bun388%
Linear368%
Neon308%
Clerk287%
Paddle247%
Remix195%
Kinde13%

For teams

What to do next

Four changes if people talk about your product online.

  • Watch GitHub issues

    Nearly half of public product questions and most bug reports go there. If your team only tracks social, you miss them.

    Developer community monitoring
  • Answer on X and Reddit

    Together they hold about 39% of public product questions, and people ask them in replies you can join.

    Reddit monitoring
  • Filter before you count

    If your name is also a word, a raw keyword count can be more than 80% noise. Remove it before you compare yourself with competitors.

    How Octolens filters mentions
  • Add Bluesky to the list

    It carries 1.9x the developer-tool conversation of Hacker News, even after a third of its volume is removed as bots.

    Bluesky monitoring

Method

How we ran it

What we did

  • We ran one search per tool per day for 30 days, September 30 to October 29, 2026. 1,423 of 1,440 tool-days completed.
  • We added 5 deeper passes that searched each large tool platform by platform, because a daily search only scores the 100 newest matches.
  • Rates come from a pool of 24,252 fully scored mentions. Octolens AI scored each one for relevance, sentiment and topic.
  • On Bluesky and Hacker News we flagged automated accounts by handle and posting pattern: status-page relays, job boards, trending feeds and accounts that mirror Hacker News. They made up 33% of relevant Bluesky mentions and under 1% on Hacker News.
  • Stack Overflow question counts come from the public Stack Exchange API. In October 2026 the site issued 6,779 post IDs in total, covering questions, answers and removed posts, so deleted posts can't change the picture.
  • We set up separate tracking for this study and picked a few of the biggest tools in each of eight categories. Nothing came from customer workspaces or from the keywords our customers track.

What the data can't tell you

  • Platform shares and daily volumes are minimums for big tools. GitHub issues, X, Reddit and news hit a limit of about 100 matches per platform per day most often.
  • We don't make claims about YouTube or LinkedIn, and podcasts are partial. Our collection skipped YouTube and podcasts on many days, and LinkedIn mentions weren't scored.
  • GitHub data means issues. It leaves out pull requests, discussions and code.
  • Sentiment and topic shares by platform include automated accounts. On Bluesky, about two in three posts tagged as bug reports came from outage-status bots, so we don't rank Bluesky for bug reports.
  • We only see public posts. Discord, Slack and questions asked inside AI assistants don't appear.
  • Our panel covers one month, so it shows where conversation sits today. The only trend on this page is Stack Overflow's own question count.
  • We make Octolens, and many developer tools use it, so some of the 48 happen to be customers. We publish sentiment by category and platform, and never by tool.

Cite this study

Use the data

Use any chart or number in articles, decks and research. Please link back.

Citation
Octolens (2026). State of Developer Conversation 2026. https://octolens.com/report/state-of-developer-conversation-2026
HTML link
<a href="https://octolens.com/report/state-of-developer-conversation-2026">Octolens State of Developer Conversation 2026</a>

Good to know

  • Each finding has its own link, like #noise.
  • Licensed under CC BY 4.0. Credit Octolens with a link.

FAQ

Frequently asked questions

Anything else? Write to hi@octolens.com.

Is there more developer conversation on Bluesky or Hacker News?

By volume, Bluesky. With bots, status feeds and Hacker News mirror accounts removed, Bluesky carried 430 relevant mentions a day about 48 developer tools, against 227 on Hacker News. That is 1.9x the volume. We counted posts, not views, and a single Hacker News front-page thread can reach more readers than hundreds of Bluesky posts.

How much of Bluesky's developer conversation is bots?

About a third. 33% of relevant Bluesky mentions came from status-page relays, job boards, trending-repo feeds and accounts that mirror Hacker News. On Hacker News the share was under 1%.

Is Stack Overflow dead?

As a place to ask new questions, close to it. Stack Overflow received 968 new public questions in October 2026, down 99.5% from 186,031 in October 2015. The archive of 24.2 million questions is still online, and we did not measure how many people read it.

How many questions does Stack Overflow get per month?

968 in October 2026 and 964 in September 2026, according to the public Stack Exchange API. The best month on record is March 2014, with 207,169. These counts leave out deleted questions.

Why did Stack Overflow decline?

Our data shows timing and can't prove a cause. October question counts fell 43% from 2015 to 2022, before ChatGPT launched. They then fell 51%, 56%, 77% and 82% in the four years after it launched.

Where do developers ask questions now?

In public, we found 46% of product questions about the 48 tools in GitHub issues, 20% on X and 19% on Reddit. Questions people ask in ChatGPT, Claude or private chat groups never appear in public data, so we can't count those.

Where do developers hang out online?

For conversation about developer tools, GitHub issues, X and Reddit carried about 72% of the relevant mentions we measured. Bluesky came fourth, with 1.9x the volume of Hacker News after we removed bots.

How did you run this study?

We searched for 48 developer tools every day from September 30 to October 29, 2026 with the Octolens search pipeline, using tracking we set up only for this study. AI scored each match for relevance, sentiment and topic. No data came from customer workspaces. The method section lists what we can and can't claim.

Can I use these numbers?

Yes. The data is licensed under CC BY 4.0. Credit Octolens and link to this page.

7-day free trial

Start listening to your customers today

No credit card required · API access on all plans · Cancel anytime