frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Learning What Will Happen Next: Predictive Coding in Hyperspace

https://blog.brojo.ai/learning-what-will-happen-next-predictive-coding-in-hyperspace/
1•bojo•28s ago•0 comments

Autogenerate Docs from GitHub

https://www.mintlify.com/blog/auto-generate-docs-from-repos
1•denssumesh•46s ago•0 comments

Teaching Al to Understand What Words Mean

https://sublius.substack.com/p/teaching-ai-to-understand-what-words-d58
1•spacebacon•1m ago•0 comments

Show HN: Goodthinking – PM skills for Claude Code

https://www.extremeclarity.ai/goodthinking
1•faizanbhat•1m ago•0 comments

Google Gemini Agent for multi-step tasks

https://gemini.google/overview/agent/
1•sheepscreek•5m ago•0 comments

Asking the raw Gemini 3.1 Pro API what kind of human it would choose to be

1•PerlBlueDot•6m ago•0 comments

Show HN: A One-Page Field Guide to Questioning

https://stefankober.github.io/essays/questioning-a-field-guide/questioning-guide/
1•kbrkbr•7m ago•0 comments

When Video Games Were Brown

https://www.youtube.com/watch?v=TTjGDkDI49I
1•CharlesW•8m ago•0 comments

Saturday Night Live criticised for 'hurtful' Tourette's sketch

https://www.bbc.co.uk/news/articles/c05v0ln6nq3o
1•mellosouls•8m ago•0 comments

Show HN: OnCallMate – AI agent for autonomous Docker incident RCA

https://github.com/ismailperim/oncallmate
1•ismailperim•9m ago•0 comments

The Pentagon's Favorite Tech Guy Is This Hawaiian Shirt-Wearing Founder

https://www.nytimes.com/2026/03/02/technology/pentagon-anduril-palmer-luckey.html
2•saikatsg•11m ago•0 comments

Fentanyl-Rewired: A 2-Azaspiro[3.3]heptane Core Preserves μ-Opioid Function

https://pubs.acs.org/doi/10.1021/acsmedchemlett.5c00672
1•PaulHoule•11m ago•0 comments

Directus Hits 500 Contributors

https://500-contributors-thingee.vercel.app/
1•haynesben•13m ago•0 comments

State of WASI support for CPython: March 2026

https://snarky.ca/state-of-wasi-support-for-cpython-march-2026/
1•ingve•13m ago•0 comments

Era of Polycrisis

https://en.wikipedia.org/wiki/Era_of_Polycrisis
2•maxwell•14m ago•0 comments

British Columbia to end time changes, adopt year-round daylight time

https://www.cbc.ca/news/canada/british-columbia/b-c-adopting-year-round-daylight-time-9.7111657
10•ireflect•14m ago•2 comments

America's Ranks of Immigrant Truckers Find a Roadblock: English Tests

https://www.wsj.com/business/logistics/americas-ranks-of-immigrant-truckers-find-a-roadblock-engl...
3•petethomas•14m ago•1 comments

Dabao Evaluation Board Risks and Challenges

https://www.crowdsupply.com/baochip/dabao#risks-challenges
1•Confiks•15m ago•0 comments

We Interviewed Our OpenClaw Agent Using a Voice Avatar

https://twitter.com/ptservlor/status/2028518999747915953
1•arbayi•15m ago•1 comments

Show HN: Exquisite.Monster – Telephone Pictionary with strangers and friends

https://exquisite.monster/
1•Swayworn•15m ago•0 comments

32 TB Gone

https://twitter.com/1v100000/status/2028319935626371137
4•worik•16m ago•2 comments

Maps and charts of Iran crisis: The global chokepoint in the Strait of Hormuz

https://www.reuters.com/graphics/IRAN-CRISIS/MAPS/znpnmelervl/
2•giuliomagnifico•17m ago•0 comments

Show HN: CrowPay – add x402 in a few lines, let AI agents pay per request

https://www.crowpay.ai/
1•ssistilli•18m ago•0 comments

Free IP Geolocation Database

https://ip66.dev/
1•todsacerdoti•21m ago•0 comments

Vykar is a fast, encrypted, deduplicated backup tool written in Rust

https://vykar.borgbase.com/
2•havaloc•23m ago•2 comments

Designing Odin's Casting Syntax

https://www.gingerbill.org/article/2026/02/23/designing-odins-casting-syntax/
1•g0xA52A2A•24m ago•0 comments

My perfect Music app doesn't exist

https://hicks.design/journal/my-perfect-music-app-doesnt-exist
2•tobr•25m ago•0 comments

The Future of AI Images

https://petergasston.co.uk/the-future-of-ai-images/
1•speckx•28m ago•0 comments

Open Source Sailing Block [video]

https://www.youtube.com/watch?v=2efa3epOGgY
1•zeapo•28m ago•0 comments

The Currency War: Why Washington's Policies Look Chaotic (But Aren't) [video]

https://www.youtube.com/watch?v=TfYB3LZf0mQ
1•csense•29m ago•1 comments
Open in hackernews

Ask HN: Self Sustaining Codebases

4•rs545837•2h ago
I've been thinking about where codebases are heading as agents write more of the code.

Right now the pipeline is: agent writes code, opens a PR, another agent or human reviews it, CI runs all 4000 tests, someone clicks merge. This is the human workflow copy-pasted onto machines. It doesn't scale.

What if a codebase could maintain itself? Agents fix bugs as they appear, add features from specs, refactor when complexity grows, update dependencies. All verified automatically, shipped continuously. Humans set policies ("anything touching auth needs my approval", "cosmetic changes auto-ship") and only see the exceptions.

The pieces that make this possible: a dependency graph of every function, class, and method in the codebase. Not files, not lines. Entities. If you know what every entity depends on, you can compute blast radius instantly, detect real conflicts (not the false ones git creates), run only the 4 tests that matter instead of 4000, and score confidence per change. High confidence ships automatically. Low confidence routes to a human.

Comments

verdverm•1h ago
The experiments with model collapse would seem to indicate that doing this will go very poorly. See also examples of claw spam for what this looks like today
rs545837•1h ago
Model collapse happens when AI output feeds back into AI training data. That's a real problem for models.

But self-sustaining codebases aren't using AI to verify AI. The verification layer is deterministic: dependency graphs, targeted test suites, blast radius computation. These are structural checks, not generative ones. The graph doesn't hallucinate. Tests either pass or they don't

The claw spam problem is what happens when you have no verification at all.

verdverm•42m ago
Model collapse happens outside of the training loop as well, at the agentic level, when left to their own devices for too long.
rs545837•1h ago
The PR was invented because humans can't be trusted to push directly to main. If automated verification is more thorough than human review (and benchmarks suggest it is), the PR becomes a receipt, not a gate.

Has anyone else been thinking along these lines? Curious if this resonates or if there are fundamental blockers I'm not seeing.