frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

What kinds of bugs does AI generate?

https://blog.detail.dev/posts/ai-bug-types/
4•drob•1h ago

Comments

pyryt•1h ago
Author here. In short:

- We launched an AI bug scanner 6 months ago. We wanted to know what the bugs that get fixed have in common.

- We clustered 1,000 bugs from 99 codebases and grouped them by failure mechanism.

- 21 recurring mechanisms cover 70% of the bugs. These same mistakes show up again and again in completely unrelated products. We seem to be writing similar bugs over and over again with agents.

- A common denominator is silence. Bugs that make it to production today are not easily ‘legible’. They don’t crash, they pass CI, they look fine. You don’t see them unless you look for them.

If you’d like to know anything else about these bugs, let us know. Happy to analyze the dataset further.

saidnooneever•52m ago
interesting data and results. it kind of looks like they are good at coding plainly no buffer overflows or such things, but as noted application logic is the trick. Things that are difficult for humans (authentication, paralelism) seems also tricky for them.

Id wonder if they are badly putting together the logic from good instructions or that the prompting was missing pieces and it followed correctly but provided some broken code due to missing requirements / details.

Ask HN: What tool are you using for A/B tests and analyzing and documenting them

1•ciwolex•30s ago•0 comments

Claude Sonnet 5

https://www.anthropic.com/news/claude-sonnet-5
8•marinesebastian•1m ago•0 comments

Building a Jax training loop for an LLM training run

https://www.gilesthomas.com/2026/06/llm-from-scratch-34a-building-a-jax-training-loop-for-an-llm-...
1•gpjt•1m ago•0 comments

Show HN: An agent harness with model autorouting and memory

1•aperi•1m ago•0 comments

Show HN: Pacific Slate – a self-hosted, model-agnostic multi-agent AI assistant

https://pacslate.com
1•badwx•1m ago•0 comments

Swiss e-ID delayed to beef up security

https://www.swissinfo.ch/eng/swiss-ai/introduction-of-e-id-delayed-for-security-reasons/91675866
1•m3drano•1m ago•0 comments

India's Unconvincing Economic Facade

https://www.aei.org/research-products/report/indias-unconvincing-economic-facade/
1•eatonphil•2m ago•0 comments

Russian politician says GTA 6 carries 'the stench of Americanism'

https://www.pcgamer.com/games/grand-theft-auto/russian-politician-says-gta-6-carries-the-stench-o...
1•HelloUsername•2m ago•0 comments

Mu: A social app with a backbone, built on European values

https://hello.mu.social/
1•modinfo•2m ago•0 comments

Eclipsa Video: HDR That Looks Right on Every Screen

https://android-developers.googleblog.com/2026/06/eclipsa-video-hdr-review.html
1•ledoge•4m ago•0 comments

Claude Sonnet 5 System Card

https://anthropic.com/claude-sonnet-5-system-card
3•adocomplete•4m ago•0 comments

Who can take on web exploit scans? The Banhammer™ can!

https://code.backwater.systems/blog/#2026-06-29T23:40:00.000Z
1•nate-gehringer•5m ago•0 comments

The DLL that was not present in memory despite not being formally unloaded 2

https://devblogs.microsoft.com/oldnewthing/20260626-00/?p=112472
1•supermatou•5m ago•0 comments

Valve explains the Steam Machine (and that pricing) [video]

https://www.youtube.com/watch?v=Y8_e3SteAgE
1•HelloUsername•5m ago•0 comments

Ask HN: Help my web browser project be better

3•roschdal•5m ago•0 comments

AgentWire: Orchestrating many Claude Code sessions via tmux and voice-control

https://github.com/dotdevdotdev/agentwire-dev
1•AdamTSaunders•5m ago•0 comments

Johnson Thermoelectric Energy Converter

https://en.wikipedia.org/wiki/Johnson_thermoelectric_energy_converter
1•simonebrunozzi•6m ago•0 comments

Prompt Injection Is Not a Chatbot Problem: How the Attack Surface Changes

https://agentsafelabs.com/blog/prompt-injection-is-not-a-chatbot-problem-how-the-attack-surface-c...
1•waqarjaved•6m ago•0 comments

Your Rust Service Isn't Leaking — It Could Be the Allocator

https://pranitha.dev/posts/rust-and-memory-allocators/
1•abhirag•6m ago•0 comments

Claude Desktop is now available on Linux (in beta)

https://twitter.com/ClaudeDevs/status/2071988881717871065
2•Recursing•6m ago•1 comments

Connect any self-hosted WordPress site to your AI

https://wpvibe.ai/start/
1•luispa•8m ago•0 comments

Comparing GLM 5.2 and Opus 4.8 implementing the same methods for the same repos

https://gist.github.com/smellslikeml/36bf4939d76f0f84d113e2ddde5e6d3c
1•mynameisfunk•10m ago•0 comments

Show HN: Deconstructing Anthropic's Coding Agent Control Model

https://www.highflame.com/blog/how-anthropic-contains-its-own-coding-agents-and-get-that-coverage...
1•sharathr•10m ago•0 comments

Apple to Adopt Next-Generation Display Color Standard

https://www.trendforce.com/presscenter/news/20260629-13124.html
1•ledoge•12m ago•0 comments

Useful LLM Prompts for Editing Your Own Technical Writing

https://techstackups.com/articles/useful-llm-prompts-for-editing-your-own-technical-writing/
1•ritzaco•13m ago•0 comments

Ford rehires 'greybeards' after AI tech fails to deliver

https://www.msn.com/en-us/news/technology/ford-rehires-human-engineers-after-ai-fails-to-match-qu...
2•ilamont•15m ago•0 comments

Shipping post-quantum cryptography to Python

https://blog.trailofbits.com/2026/06/30/shipping-post-quantum-cryptography-to-python/
1•logickkk1•17m ago•0 comments

Forget Work. Passive Income Is the New American Dream

https://www.wsj.com/lifestyle/careers/passive-income-dreams-2e67ee5c
1•apparent•19m ago•0 comments

California Is Chasing Wealth That Has Feet

https://progressandpoverty.substack.com/p/california-is-chasing-wealth-that
3•slackpad•20m ago•2 comments

Millions take omega-3 fish oil for brain health but a new study found no benefit

https://www.sciencedaily.com/releases/2026/06/260622014300.htm
2•eustoria•22m ago•0 comments