24:I[50237,["/_next/static/chunks/0db4v-xi1q-jq.js","/_next/static/chunks/02-tqo9.1..63.js","/_next/static/chunks/0d3shmwh5_nmn.js","/_next/static/chunks/12wcqrolhhiqu.js"],"Posts"] 25:T421,Inspired by Karpathy's HackerTube, I wondered what was buried in "What are you working on?" thread comments. I wanted to see what changed, and who was still at it years later. So, I built this site to answer some of those questions.

A couple of things you can do: type in your idea and see who already tried it, look up your username and see what you were working on in 2020, see who’s been working on the same thing for a decade, or what founders posted before they had a company.

I built it using Claude Code. Claude Sonnet read all 20,982 top-level replies and pulled out a name, description, topic and how AI was involved. One finding is that AI went from 1% of answers in 2009–11 to 33% this year.

A few limits of the result: labels are model output, and I only spot-checked them. The HN thread only went monthly in 2024, so early years are thin. If I’ve mislabeled your project, or you’d rather not be included, please tell me.

If you’ve been posting here for years, I’d love to hear what happened to those projects.26:T4e4,Hey HN community, I built MergeAtlas - a better way to review PRs when bulk of the code is written by AI agents. It is a skill for your coding agent plus a small python script.

MergeAtlas was built from a simple premise - When AI agents are writing the code, there's no need to review every single line of code. Instead it is better to understand the intent, important lines, and the blast radius.

MergeAtlas helps you see the intents behind file changes, understand why each file was changed, and PR's scope all before you read a single line of code. When you are ready to read the code, MergeAtlas singles out the key lines from each file, so you are only focused on the most important lines.

It goes further than GitHub's Files Changed tab as it also determines what files outside of the PR are affected by the changes.

I tuned the skill till I was satisfied with the result in terms of structure, explanations, and verbosity.

There's a live demo on the website based on an actual PR from excalidraw. Play around with it, and any feedback is welcome.

Github: https://github.com/ismayilkarimli/MergeAtlas

NIST Adversarial Machine Learning: Public Draft for Comments

Anthropic launches new Claude Fable and Mythos 5.2 models

STM32 Memory Protection in Practice

A weakness could reduce chemotherapy reliance in triple-negative breast cancer

Street homeless for a year. Here is a public log (work in progress)

QUIC: A general-purpose transport layer network protocol

Show HN: Search 17 years of "What are you working on?"

Testing Claude on real-world electronics design

Claude Fable 5.2

In 2023 a Danish customer lost power for 30 minutes on average. An Estonian, 419

Claude Code memories that save the fix, not just a note

Ask HN: Is AI involved in on-the-field sports strategies?

Claude Fable 5.2 and Mythos 5.2

Automake 1.20 Brings New Features, to Re-Establish GNU as Leading Build System

B.C. businesses push back against 'AI slop' with pen-and-paper ads

AMD PerfOpt Performance Boosting Code Tweaked for Linux 7.4-Rc2

Show HN: MergeAtlas – A better way to review PRs

Do We Still Have to Denormalize? Have Your Schema and Search It Too

Synaptics sends cease-and-desist order to open-source DisplayLink hobbyist

Why Gmx.com?

GrapheneOS myths I busted after installing it on my Pixel

Vodafone Deutschland to axe 2,300 jobs as it steps up cost-cutting efforts

Lessons learned when hiring engineers in the age of AI

GNU Groff 1.25.0 Arrives with Typesetting Fixes and Refinements

E-means: 10x faster k-means at equal loss

Hackers exploit SSHole OpenSSH flaw to breach Linux systems

Upside-Down-Ternet (2006)

Security Vocabulary for Maintainers

Encyclopedia of Triangle Centers

PolinRider: North Korea-linked worm forges developers' commits in 7,609 repos

NIST Adversarial Machine Learning: Public Draft for Comments

Anthropic launches new Claude Fable and Mythos 5.2 models

STM32 Memory Protection in Practice

A weakness could reduce chemotherapy reliance in triple-negative breast cancer

Street homeless for a year. Here is a public log (work in progress)

QUIC: A general-purpose transport layer network protocol

Show HN: Search 17 years of "What are you working on?"

Testing Claude on real-world electronics design

Claude Fable 5.2

In 2023 a Danish customer lost power for 30 minutes on average. An Estonian, 419

Claude Code memories that save the fix, not just a note

Ask HN: Is AI involved in on-the-field sports strategies?

Claude Fable 5.2 and Mythos 5.2

Automake 1.20 Brings New Features, to Re-Establish GNU as Leading Build System

B.C. businesses push back against 'AI slop' with pen-and-paper ads

AMD PerfOpt Performance Boosting Code Tweaked for Linux 7.4-Rc2

Show HN: MergeAtlas – A better way to review PRs

Do We Still Have to Denormalize? Have Your Schema and Search It Too

Synaptics sends cease-and-desist order to open-source DisplayLink hobbyist

Why Gmx.com?

GrapheneOS myths I busted after installing it on my Pixel

Vodafone Deutschland to axe 2,300 jobs as it steps up cost-cutting efforts

Lessons learned when hiring engineers in the age of AI

GNU Groff 1.25.0 Arrives with Typesetting Fixes and Refinements

E-means: 10x faster k-means at equal loss

Hackers exploit SSHole OpenSSH flaw to breach Linux systems

Upside-Down-Ternet (2006)

Security Vocabulary for Maintainers

Encyclopedia of Triangle Centers

PolinRider: North Korea-linked worm forges developers' commits in 7,609 repos

Ask HN: Is AI involved in on-the-field sports strategies?

Comments