frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Show HN: We Built 450 Modular Agent Skills for Medical Research

https://github.com/aipoch/medical-research-skills
1•The_resa•2m ago•0 comments

Show HN: Smux – split terminals for AI agents

https://github.com/gergomiklos/smux
1•gregolo•2m ago•0 comments

4.0M tokens for Next.js vs. 2.5M tokens for Wasp: same app, same prompt

https://wasp.sh/blog/2026/03/26/nextjs-vs-wasp-40-percent-less-tokens-same-app
1•matijash•5m ago•0 comments

Ask HN: Why do you flag a service when it goes down?

1•haebom•7m ago•0 comments

Vibe coding could mark the end of the App Store review process as we know it

https://9to5mac.com/2026/03/29/vibe-coding-developers-report-long-app-store-review-queues/
1•thm•8m ago•0 comments

The case for a universal basic income in the era of A.I

https://www.americamagazine.org/short-take/2026/03/23/universal-basic-income-ai/
1•robtherobber•10m ago•0 comments

Galton's Law of Mediocrity: Why Large Language Models Regress to the Mean

https://arxiv.org/abs/2509.25767
2•camillomiller•12m ago•0 comments

Stripe stole $85k and closed our account

5•MelkerWendelbo•17m ago•4 comments

Apple nailed AI by doing fucking nothing lol

https://xcancel.com/cryptopunk7213/status/2038351931589193953?s=20
3•doener•21m ago•0 comments

Stripe Dashboard Down

1•HPMOR•21m ago•0 comments

Stripe Is Down

https://dashboard.stripe.com/login
8•tompccs•22m ago•0 comments

In Praise of Plasma TVs

https://hackaday.com/2025/11/18/in-praise-of-plasma-tvs/
1•jruohonen•23m ago•0 comments

Show HN: AnyVali – Validation library that works the same across 10 languages

https://anyvali.com/
1•mrinc•25m ago•0 comments

GPL-compliant reasonable legal notices and author attributions

https://www.fsf.org/blogs/community/gpl-compliant-legal-notices-author-attributions
2•Tomte•27m ago•0 comments

You're right to be anxious about AI: This is how much we are building

https://www.dumky.net/posts/youre-right-to-be-anxious-about-ai-this-is-how-much-we-are-building/
3•dmkii•38m ago•3 comments

Mathematical methods and human thought in the age of AI

https://terrytao.wordpress.com/2026/03/29/mathematical-methods-and-human-thought-in-the-age-of-ai/
2•jjgreen•40m ago•0 comments

Swift SDK for Android

https://www.swift.org/documentation/articles/swift-sdk-for-android-getting-started.html
2•devy•46m ago•0 comments

Polygraphs have major flaws. Are there better options?

https://undark.org/2026/03/25/lie-detection-polygraph-accuracy/
3•Tijana329•49m ago•0 comments

Stripe Is Down

https://downdetector.fr/en/status/stripe/
6•pinter69•51m ago•0 comments

Show HN: IsDisposable – Open-source disposable email detection (160K+ domains)

https://www.npmjs.com/package/@isdisposable/js
2•junaidshaukat•52m ago•0 comments

Mistral raises $830M to build Nvidia-powered AI centres in Europe

https://www.ft.com/content/229f4f59-d518-4e00-abd6-5a5b727cd2aa
3•macleginn•52m ago•2 comments

Show HN: I'm a Happy Engineer [video]

https://www.youtube.com/watch?v=f1a_MRLibqU
2•denysvitali•55m ago•0 comments

How to Survive in the Tech industry in 2026

https://blog.phuaxueyong.com/post/2026-03-23-how-to-survive-tech-in-2026/
8•xueyongg•1h ago•2 comments

How Can Universities Value-Add Their Alumni?

https://blog.phuaxueyong.com/post/2025-06-27-university-role-in-alumni-engagement/
2•xueyongg•1h ago•0 comments

The CTO's Burden: Building What the World Doesn't See

https://blog.phuaxueyong.com/post/2025-04-29-questions-for-cto/
2•xueyongg•1h ago•0 comments

Show HN: Travel app that replaces trip research with a 30s briefing (TestFlight)

https://globallybased.com
2•ilyagruzhevski•1h ago•3 comments

Sad Story of Soviet Compact Disc Players

https://sovietrock.com/mediums/cd/sad-story-of-soviet-compact-disc-players/
2•thenthenthen•1h ago•0 comments

Credential Broker for Agents (CB4A)

https://datatracker.ietf.org/doc/draft-hartman-credential-broker-4-agents/
1•jruohonen•1h ago•0 comments

We tricked 1M+ bots and hackers with our honeypot

https://github.com/BlessedRebuS/Krawl
2•blessedrebus•1h ago•1 comments

Every Package You Install Can Read Your Secrets

https://www.eliranturgeman.com/2026/03/28/supply-chain-attacks/
1•gsky•1h ago•1 comments
Open in hackernews

Hamilton-Jacobi-Bellman Equation: Reinforcement Learning and Diffusion Models

https://dani2442.github.io/posts/continuous-rl/
22•sebzuddas•1h ago

Comments

measurablefunc•1h ago
It's not clear or obvious why continuous semantics should be applicable on a digital computer. This might seem like nitpicking but it's not, there is a fundamental issue that is always swept under the rug in these kinds of analysis which is about reconciling finitary arithmetic over bit strings & the analytical equations which only work w/ infinite precision over the real or complex numbers as they are usually defined (equivalence classes of cauchy sequences or dedekind cuts).

There are no dedekind cuts or cauchy sequences on digital computers so the fact that the analytical equations map to algorithms at all is very non-obvious.

jampekka•41m ago
Continuous formulations are used with digital computers all the time. Limited precision of floats sometimes causes numerical instability for some algorithms, but usually these are fixable with different (sometimes less efficient) implementations.

Discretizing e.g. time or space is perhaps a bigger issue, but the issues are usually well understood and mitigated by e.g. advanced numerical integration schemes, discrete-continuous formulations or just cranking up the discretization resolution.

Analytical tools for discrete formulations are usually a lot less developed and don't as easily admit closed-form solutions.

phreeza•37m ago
Doesn't continuous time basically mean "this is what we expect for sufficiently small time steps"? Very similar to how one would for example take the first order Taylor dynamics and use them for "sufficiently small" perturbations from equilibrium. Is there any other magic to continuous time systems that one would not expect to be solved by sufficiently small time steps?
measurablefunc•8m ago
You should look into condition numbers & how that applies to numerical stability of discretized optimization. If you take a continuous formulation & naively discretize you might get lucky & get a convergent & stable implementation but more often than not you will end up w/ subtle bugs & instabilities for ill-conditioned initial conditions.
nareyko•43m ago
One interesting connection is that many production AI systems don't explicitly implement RL frameworks, but still behave like RL systems.

You still have: state -> user context action -> model output reward -> engagement or success metric

Once that loop exists, optimization dynamics start to look very similar.

Cloudly•41m ago
Ever since the control bug bit me in my EE undergrad years I am happy to see how useful the knowledge remains. Of course the underlying math of optimization remains general but the direct applications of control theory made it much more appetizing for me to struggle through.