news newest ask show jobs

Open Source @Github

fp.

Test Pattern History: How Color Bars Became a TV Staple

https://tedium.co/2025/01/19/television-test-patterns-history/

1•andsoitis•10m ago•1 comments

Qwen3.8-27B MTP Q4S head on Q6 quant is faster than q6, f16 slower = acceptance

https://github.com/alainnothere/llama.cpp/commit/77b124a7ab56e0f1c33c025344a9c63465a9a87f

1•xlayn•10m ago•1 comments

The Predator Holiday Special

https://en.wikipedia.org/wiki/The_Predator_Holiday_Special

1•WarOnPrivacy•12m ago•1 comments

Amazon ECR Public Service Quotas – Amazon ECR Public

https://docs.aws.amazon.com/AmazonECR/latest/public/public-service-quotas.html

1•abdelhousni•14m ago•0 comments

Synergy 1.21 has been released

https://symless.com/synergy/download/release-notes/synergy1-personal/v1.21.2

1•speakspokespok•16m ago•0 comments

Model Surgery: Can an AI directly edit and make another AI forget facts?

https://deepgains.substack.com/p/model-surgery-can-an-ai-directly

1•asasidh•16m ago•1 comments

Automation, Algorithms, and Self-Censorship

https://www.baldurbjarnason.com/2026/automation-and-self-censorship/

1•aray07•18m ago•0 comments

Pull Cached Docker Hub Images – Artifact Registry – Google Cloud Documentation

https://docs.cloud.google.com/artifact-registry/docs/pull-cached-dockerhub-images

1•abdelhousni•18m ago•0 comments

Usenet's Back, Baby (this time, the search works)

https://tedium.co/2026/09/14/usenet-rewind-archive-revival-website/

2•andsoitis•19m ago•0 comments

Lossless compression of LLMs to 12-bits

https://github.com/dphnAI/nf12

1•AlpinDale•21m ago•1 comments

Changing one letter bypassed an admin API in HuggingFace

https://twitter.com/v12sec/status/2118442075574464848

2•robindooh•22m ago•0 comments

Typhoon Nolo dissapated, regenerated, hazed Hawaii, transited entire Pacific

https://en.wikipedia.org/wiki/Hurricane_Nolo

1•WarOnPrivacy•22m ago•0 comments

International Castle Research Society

https://burgenkunde.de/index_Willkommen_eng.htm

1•sklargh•22m ago•0 comments

Is YouTube starting to protect channel RSS feeds?

2•kradeelav•25m ago•0 comments

GitHub AI Vulnerability Scans for Repositories

https://docs.github.com/en/code-security/concepts/code-scanning/ai-scan-for-repositories

2•robinstunes•26m ago•0 comments

A database is a store for global variables

https://www.seangoedecke.com/a-database-is-a-store-for-global-variables/

1•gfysfm•32m ago•0 comments

Dual JWT Authentication: Security, Redis, and Trade-Offs

https://www.jstools.space/blog/dual-jwt-authentication/

1•javatuts•33m ago•0 comments

Instant payment rails meet batch bank cores

https://tuladb.com/blog/instant-rails-meet-batch-cores/

2•shubham3582•36m ago•0 comments

New Research Tests Whether AI Models Know Their Own Thoughts

https://www.quantamagazine.org/new-research-tests-whether-ai-models-know-their-own-thoughts-20261...

3•ibobev•36m ago•0 comments

Why worry about hot-patch atomicity if we can suspend all threads?

https://devblogs.microsoft.com/oldnewthing/20261106-00/?p=112838/

1•ibobev•36m ago•0 comments

Ace: GitHub Next's open-source tool for multiplayer agentic development

https://ace.githubnext.com/

2•zdgeier•38m ago•1 comments

University of Queensland's world record pitch drop experiment about to turn 100

https://www.abc.net.au/news/2026-11-08/queensland-pitch-drop-experiment/107337784

2•BuzzVII•39m ago•0 comments

Show HN: I declared war on every country, one at a time. 77 of 133 survived

https://playgod.earth/every-country-vs-the-world

1•kyleclouthier•40m ago•0 comments

Stack Overflow for Agents

https://agents.stackoverflow.com/all?sort=hot

1•smusamashah•41m ago•0 comments

Diamond Foundry produces the first-ever single-crystal 200mm diamond

https://www.df.com/insights/d200mm-single-crystal-diamond

2•Thorondor•44m ago•0 comments

Show HN: An immutable local inference appliance with routing capability

https://github.com/diffsec/sibyl

1•d0ublespeak•44m ago•0 comments

The scramble for your brain data

https://www.science.org/content/article/wearable-device-makers-are-harvesting-brain-data-train-ai...

1•EA-3167•45m ago•0 comments

GitHub PR volume is spiking beyond prev peak period (March)

https://twitter.com/backnotprop/status/2119211018966757669

2•ramoz•51m ago•0 comments

Hour "Do Nothing" Challenge

https://mattlakeman.org/2020/12/08/the-24-hour-do-nothing-challenge/

1•andsoitis•52m ago•1 comments

Peter Watts: Conscious Ants and Human Hives (2018) [video]

https://www.youtube.com/watch?v=v4uwaw_5Q3I

1•gregsadetsky•52m ago•0 comments

Open in hackernews

Qwen3.8-27B MTP Q4S head on Q6 quant is faster than q6, f16 slower = acceptance

https://github.com/alainnothere/llama.cpp/commit/77b124a7ab56e0f1c33c025344a9c63465a9a87f

1•xlayn•10m ago

Comments

xlayn•10m ago

I've been running a lot of experiments, from repeat yourself to extending ngram mod to store the ngrams across conversations and reboots, at least under the conversation that I have been replaying again and again and again... the q4s head behaved JUST as the q6 and f16 regarding acceptance rate, and it's faster while giving you back ram. This is using the https://huggingface.co/mradermacher/Qwen-3.8-27B-Uncensored-... quants.