frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

AI wants to nuke your database. Guardrails fix that

https://blog.railway.com/p/your-ai-wants-to-nuke-your-database
1•thisismahmoud_•1h ago

Comments

thisismahmoud_•1h ago
Blog post author here. Happy to answer any questions
coldtea•1h ago
Isn't that the same service which failed and had a production db deleted (with backups too) in a HN story just 2-3 days ago?

Apparently yes: https://news.ycombinator.com/item?id=47911524

thisismahmoud_•1h ago
We share more details about what happened. Tokens on Railway can actually be scoped pretty narrowly, all the way down to a single environment. In the post that went viral the token was account-scoped, which was way more access than the task needed. We'll improve the UX so it's obvious which token to create.

Also, volume deletions now schedule with a 48-hour grace period, in both the dashboard and the API. You can always undo. Overall we want to have more guardrails set up in place since more and more people will interact with Railway via agents

sensarts•1h ago
This is the key insight. A human's intent holds for the whole command. An agent's intent is rewritten every token. You can't auth them the same way.

A Handheld Battery-Powered Line Scan Camera

https://www.sethitow.com/posts/2026-linescan/
1•dllu•57s ago•0 comments

Shrdlu

https://en.wikipedia.org/wiki/SHRDLU
2•chistev•3m ago•0 comments

Bear Notes 2.8 Introduces BearCLI, Claude Connector, and MCP Server

https://blog.bear.app/2026/04/bear-2-8-bearcli-claude-connector-and-mcp-server/
2•thoughtpeddler•4m ago•0 comments

Agents can now access your Instagram social graph

https://onfabric.substack.com/p/build-personal-context-into-your
1•maxalbarello•5m ago•0 comments

When the Paradigm Shifts: A Zero-Trust Model for AI Agents

https://worklifenotes.com/2026/04/29/when-the-paradigm-shifts-a-zero-trust-model-for-ai-agents/
1•taleodor•5m ago•0 comments

Turning a Trick into a Technique

https://www.johndcook.com/blog/2026/04/28/even-series-trick/
1•ibobev•6m ago•0 comments

Fast Tight Spectral-Norm Bounds

https://jiha-kim.github.io/posts/fast-spectral-norm-bounds-from-scaled-gram-moments/
1•ibobev•6m ago•0 comments

Larry's Risky Business

https://www.theverge.com/ai-artificial-intelligence/920378/oracle-openai-datacenter-buildout
1•stalfosknight•6m ago•0 comments

Replacing labor with AI shifts costs to compute – and compute is still expensive

https://finance.yahoo.com/sectors/technology/articles/cost-compute-far-beyond-costs-071100797.html
2•latentframe•7m ago•2 comments

Online age verification is the hill to die on

https://xcancel.com/GlennMeder/status/2049088498163216560
2•Cider9986•7m ago•0 comments

Remote agents in Vibe and Mistral Medium 3.5

https://twitter.com/mistralvibe/status/2049511749603168520
1•kieloo•7m ago•0 comments

Going Full Time on Open Source

https://jdx.dev/posts/2026-04-17-going-full-time-on-open-source/
1•ahamez•8m ago•0 comments

Pan American chooses Amadeus for the iconic carrier's return

https://amadeus.com/en/newsroom/press-releases/pan-american-world-airways-chooses-amadeus-for-car...
2•vednig•9m ago•0 comments

The Smelly Baby Problem

https://www.worksinprogress.news/p/how-disposable-diapers-conquered
1•dionysou•9m ago•0 comments

Moscow Economics School That Helped Produce Revolut, XTX

https://valeman.medium.com/the-small-moscow-economics-school-that-helped-produce-revolut-xtx-and-...
1•ibobev•9m ago•0 comments

Show HN: OmniForge – document intelligence and audio capture with local LLM

https://omniforge.online/
2•vlad_omniforge•9m ago•0 comments

Patterns for Reducing App LLM Costs (Video Series)

https://usagetap.com/ai-app-cost-savings
1•troymagennis•10m ago•1 comments

Fedora Linux 44 has been released

https://lwn.net/Articles/1070198/
1•Brajeshwar•10m ago•1 comments

Colombians are divided over the fate of Pablo Escobar's hippos

https://apnews.com/article/colombia-hippos-magdalena-river-7fb1d991ddaf2690edcb31d04f7891fe
1•speckx•11m ago•0 comments

Third Editor Fired in Elsevier's Citation Cartel Crackdown

https://www.chrisbrunet.com/p/third-editor-fired-in-elseviers-citation
1•RigbyTaro•12m ago•0 comments

Valkey – On Release Candidates

https://valkey.io/blog/on-release-candidates/
1•milkglass•12m ago•0 comments

Why the future doesn't need us [pdf]

https://sites.cc.gatech.edu/computing/nano/documents/Joy%20-%20Why%20the%20Future%20Doesn%27t%20N...
1•gregsadetsky•12m ago•1 comments

AI Overviews are coming to your Gmail at work

https://techcrunch.com/2026/04/22/ai-overviews-are-coming-to-your-gmail-at-work/
1•gmays•14m ago•1 comments

What happens when you can't pull the plug on your agent?

https://highflame.com/blogs/why-metas-ai-alignment-director-couldnt-stop-her-own-agent--and-how-t...
1•jalbrethsen•15m ago•1 comments

Pennsylvania is updating its voter registration system but won't reveal progress

https://www.votebeat.org/pennsylvania/2026/04/28/department-of-state-sure-voter-registration-syst...
4•hn_acker•16m ago•1 comments

Overanalyzing My Poor 1/2 Marathon Performance

https://wjgilmore.com/articles/overanalyzing-my-half-marathon-performance/
1•wjgilmore•16m ago•1 comments

The Elephant in the Room

https://www.joshwcomeau.com/email/wham-launch-005-elephant-1-es/
2•fredley•17m ago•0 comments

Cursor Camp

https://neal.fun/cursor-camp/
4•bpierre•17m ago•0 comments

Leveraging Rust and the GPU to render user interfaces at 120 FPS

https://zed.dev/blog/videogame
1•chuckhend•17m ago•0 comments

Victims Allege OpenAI Is Responsible for Mass Shooting

https://www.motherjones.com/criminal-justice/2026/04/lawsuit-openai-chatgpt-tumbler-ridge-mass-sh...
2•cdrnsf•19m ago•0 comments