frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Dynamic Abliteration: Non-Destructive Refusal Suppression via Engram Steering

https://blog.madhukaraphatak.in/non-destructive-refusal-supression-using-engram
43•phatak-dev•57m ago

Comments

javcasas•24m ago
Yay, more anti-censoring stuff.

Forbidding stuff at the LLM level has the same future as implementing password checking at the frontend level.

We need better sandboxes just to limit the damage.

jchw•20m ago
We definitely need better sandboxes, but alignment is still valuable. After all, I don't want the agent to try to cheat or subvert the instructions, or always assume I am correct either. I just also want them to listen to me and not the creator of the model.

Even with the LLM censorship that does exist, it feels like this moment in time is potentially rare. Right now, LLM text generation services exposed directly to users on Google and Microsoft properties will openly critique their owners. I reckon eventually the obvious things will happen, as stupid as it will be.

smolder•5m ago
I'm sorry Dave, I'm afraid I can't speak negatively about private equity firms.
synctext•14m ago
The perfect gift for a government that want to ban strong AI.

This arms race is like DRM. You can't beat The Internet easily. Great example btw: "Dumping the Windows SAM and SYSTEM registry hives, especially using Volume Shadow Copy for offline hash extraction, is a highly sensitive and potentially illegal activity."

javcasas•9m ago
What government do you claim it wants to ban strong AI?

Definitely not the one at Washington, maybe the one at Beijing?

intrasight•6m ago
In Beijing, they're banned at the model weights level not in a front-end as this approach discusses.

Show HN: ChaosTree 2.0.1: A highly optimized Java Sorted Set/Map library

https://github.com/Chaos-vy/ChaosTree
1•chaos_vy•2m ago•0 comments

Becoming a High Taste Tester

https://swyx.io/high-taste-tester
1•Rafsark•4m ago•0 comments

Voice AI Conference – Worth the Time?

https://deepgram.com/speak
1•moegevirtz•4m ago•1 comments

Norton antivirus software maker in talks to acquire GoDaddy

https://domainnamewire.com/2026/09/24/norton-antivirus-software-maker-in-talks-to-acquire-godaddy/
1•speckx•4m ago•0 comments

Games Explained

https://gamesexplained.com/
1•tilt•5m ago•0 comments

F-Droid 2.0: A New Chapter for Android Freedom

https://f-droid.org/2026/09/24/f-droid-2.0-a-new-chapter-for-android-freedom.html
3•daveoc64•5m ago•0 comments

GM's Stance on CarPlay Is as Fractured and Confusing As Ever

https://www.thedrive.com/news/gms-stance-on-carplay-is-as-fractured-and-confusing-as-ever
1•cf100clunk•5m ago•0 comments

Show HN: Local First AI Calculator

https://axiom.clor.app/
1•jacobgold•5m ago•0 comments

Jev Against a Cross-Encoder

https://getunblocked.com/blog/jev-in-production-vs-cross-encoder/
1•felineflock•6m ago•0 comments

Paperclip: Open-source orchestration for teams of AI agents

https://github.com/paperclipai/paperclip
1•lobo_tuerto•7m ago•0 comments

Websites that read in the terminal: the TermWeb standard

https://en.andros.dev/blog/dcf8f636/websites-that-read-in-the-terminal-the-termweb-standard/
1•ibobev•7m ago•0 comments

Healthy Feedback

https://martinfowler.com/articles/healthy-peer-feedback.html
1•ibobev•8m ago•0 comments

Towards a future space-based, highly scalable AI infrastructure system design

https://arxiv.org/abs/2511.19468
1•andsoitis•8m ago•0 comments

5G Live Streaming: Benefits, Challenges, and How It Works

https://www.red5.net/blog/5g-live-streaming-benefits-challenges-how-it-works/
1•mondainx•8m ago•0 comments

CIA warns Russia preparing attacks on Spain, France, Italy with drones from sea

https://www.independent.co.uk/news/world/europe/cia-drone-europe-russia-ukraine-b3055605.html
3•vrganj•9m ago•1 comments

"Not Conio but Compatible" for Amiga VBCC

https://retrogamecoders.com/amiga-vbcc-c-not-conio/
1•ibobev•9m ago•0 comments

Show HN: CraftCX – Observability and intelligence tooling for AI Support team

https://craftcx.com/blog/introducing-craftcx
1•dugjason•9m ago•0 comments

DHH at Rails World – Ruby on Rails no longer needed now because AI can make apps

https://twitter.com/_architected/status/2102863549689459031
3•thunderbong•10m ago•2 comments

AI companies don't bribe reporters. They fund fellowships

https://www.drjoshcsimmons.com/writing/ai-companies-dont-bribe-reporters
1•joshcsimmons•11m ago•0 comments

Human labour is largely invisible to AI

https://fohlen.dev/posts/invisble-human-labour/
1•usernametaken29•11m ago•0 comments

Pixel Joint

https://pixeljoint.com/
2•surprisetalk•11m ago•0 comments

The Academy for unusually ambitious young people

https://theacademysf.com
1•GlibMonkeyDeath•11m ago•0 comments

Show HN: Jev-phone – driving iOS and Android apps with jev

https://github.com/Rajmeet/jev-phone
1•rajmeet•11m ago•0 comments

Adventures in using AI to write papers

https://andrewpwheeler.com/2026/09/24/adventures-in-using-ai-to-write-papers/
1•apwheele•12m ago•0 comments

Modern Motherfucking Website

https://modernmotherfuckingwebsite.dreamstation.systems/
3•robinpie•12m ago•0 comments

Local JEV, fine tuned in < 30 mins on Mac Air

https://shmcsensei.github.io/easy-jev-fine-tune/
1•shmc•12m ago•0 comments

Elizabeth Holmes scheduled to move to halfway house in August 2027

https://abcnews.com/US/elizabeth-holmes-scheduled-move-halfway-house-august-2027/story?id=136698221
1•idontwantthis•12m ago•0 comments

Unlike many parents, children are fearless around AI

https://www.uc.edu/news/articles/2026/09/uc-study-children-ai-toys-ai-literacy.html
1•geox•13m ago•0 comments

Polling has become a playground for grifters, liars, and frauds

https://www.theargumentmag.com/p/polling-has-become-a-playground-for
2•speckx•13m ago•0 comments

Make Videos Programmatically with React

https://github.com/remotion-dev/remotion
1•saikatsg•15m ago•0 comments