frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Show HN: Fluidic Network Grid – A branchless control plane in Jax

https://github.com/PJHkorea/Fluidic_Network_Grid
1•PJHkorea•1m ago•0 comments

Imud – Gpsd, but for IMUs

https://github.com/richcreations/imud
1•richcreations•4m ago•0 comments

Covid-19 Misinformation

https://en.wikipedia.org/wiki/COVID-19_misinformation
2•num42•9m ago•0 comments

Foldkit

https://foldkit.dev/
2•handfuloflight•10m ago•0 comments

Video-use: Edit videos with coding agents

https://github.com/browser-use/video-use
3•nateb2022•11m ago•0 comments

Five US tech giants' hidden debts soar to $1.65T on opaque AI funding

https://asia.nikkei.com/business/technology/five-us-tech-giants-hidden-debts-soar-to-1.65tn-on-op...
6•NordStreamYacht•12m ago•0 comments

Salami Slicing Tactics

https://en.wikipedia.org/wiki/Salami_slicing_tactics
2•num42•12m ago•0 comments

Running Doom on Our Custom CPU and Going Viral

https://www.armaangomes.com/blogs/doom/
3•arghunter•14m ago•0 comments

A Koi Pond Mosaic Made from 10 Pounds of 3D Printer Waste

https://www.instructables.com/A-Koi-Pond-Mosaic-Made-From-10-Pounds-of-3D-Printe/
3•sudo_cowsay•16m ago•0 comments

Jane Street: Incremental

https://github.com/janestreet/incremental
2•handfuloflight•17m ago•0 comments

Windows Performance Exploitation

https://lagtency.com/2026/06/25/win-perf/
2•proxylatte•18m ago•0 comments

VTubing: How a Japanese Phenomenon Is Going Worldwide

https://www.tokyodev.com/articles/vtubing-how-a-japanese-phenomenon-is-going-worldwide
3•pwim•25m ago•0 comments

Emoticons

https://en.wikipedia.org/wiki/List_of_emoticons
3•audreyfei•29m ago•0 comments

How did I publish a paper in ACMKDD as an undergraduate

https://medium.com/@JarrettYe/how-did-i-publish-a-paper-in-acmkdd-as-an-undergraduate-c0199baddf31
2•ktimespi•34m ago•1 comments

Writers need new strategies to stand up to AI

https://www.thedial.world/articles/news/copyright-law-ai-intellectual-property
5•colinprince•35m ago•0 comments

Gary Marcus: US won't "win" AI war with China. What we should do instead.

https://garymarcus.substack.com/p/china-has-all-but-caught-up-the-us
3•giardini•38m ago•0 comments

Scientists Visualize the Complex, Dynamic World Inside a Human Cell

https://mymodernmet.com/eukaryotic-cell-digizyme/
2•o4c•40m ago•0 comments

US judge approves Anthropic's $1.5B settlement of copyright lawsuit

https://finance.yahoo.com/technology/ai/articles/us-judge-approves-anthropics-1-204851948.html
2•lilerjee•41m ago•0 comments

Magnet Fishing on a Military Base [video]

https://www.youtube.com/watch?v=l0uMGoW1KPM
2•xvxvx•42m ago•1 comments

New spinning drone hides in plain sight

https://news.northwestern.edu/stories/2026/07/new-spinning-drone-hides-in-plain-sight
2•o4c•45m ago•0 comments

A Practical Guide to Calculating, Pricing, and Building Recurring

https://ebookbizai.com/store/maggiebookhub/home
2•bokeke1•46m ago•0 comments

Building Agentic Workflows in Python with LangGraph

https://machinelearningmastery.com/building-agentic-workflows-in-python-with-langgraph/
3•eigenBasis•46m ago•0 comments

Mevion: Low-Cost Open-Source Data Collection System for Dual-Arm Manipulation

https://haraduka.github.io/mevion-hardware/
2•ilreb•48m ago•0 comments

AI Is Changing the Economics of Native Apps

https://medium.com/@robsandhu/ai-is-changing-the-economics-of-native-apps-e81c7f325948
3•rsandhu•49m ago•0 comments

DepthART: Scaling Foundation Monocular Depth to Tiny Models

https://xuefeng-cvr.github.io/DepthART/
2•ilreb•49m ago•0 comments

Is surveillance risk chilling your online speech?

25•Webstir•51m ago•20 comments

Show HN: Turn casual photos into professional headshots with AI

https://portraify.app
6•camel2243•55m ago•2 comments

Hugging Face warns an autonomous AI agent hacked its network

https://www.bleepingcomputer.com/news/security/hugging-face-breach-autonomous-ai-agent-system-int...
4•sbulaev•1h ago•0 comments

Invisible Details

https://invisibledetails.com
2•handfuloflight•1h ago•0 comments

Snake-and-commits: Turn your GitHub contribution graph into a real game of Snake

https://github.com/dahan8473/snake-and-commits
2•thunderbong•1h ago•0 comments
Open in hackernews

Controlling Reasoning Effort in LLMs

https://magazine.sebastianraschka.com/p/controlling-reasoning-effort-in-llms
63•ibobev•13h ago

Comments

simonw•9h ago
I'm amused by how the whole reasoning model thing feels like a formalization of the old "think step by step" prompting hack, which was discovered against GPT-3 two years after that model was first released.

My favorite trick for controlling the reasoning level is the hack where you look at the output token stream and spot the token for "the model has concluded reasoning"... and then replace that with the tokens for "wait, but" and force it to keep going!

embedding-shape•6h ago
> My favorite trick for controlling the reasoning level is the hack

First described here I think, specifically for Qwen models: https://arxiv.org/abs/2501.19393, although it seems unclear how much it actually improves the final response unless you create a new fine-tuned model with this in mind, and also if it's possible to replicate for other models. Still, got inspired and gonna give it a try with DiffusionGemma although slightly differently

sva_•9h ago
I recommend his book, "Build a Reasoning Model (From Scratch)", which is also linked in the article.

https://sebastianraschka.com/books/#build-a-reasoning-model-...

razorbeamz•3h ago
LLMs don't actually reason, they just create the illusion of reasoning by repeating things over and over.
cyanydeez•2h ago
mmm, practically, llamacpp solved this with reasoning budget for tokens and a customized message.

ive tailored my local models to match agent with a message that pushes to use subagents and context compression.

it works pretty smoothly if theres proper vector and scope to the project. theres probably a way to also get it to record memories but ive not seen a memory system that includes spatial type reasoning which would create memories associated with file paths down to AST graphlike leaves.

then we could include remembering details about the path.

llamacpp also has a header for setting these that an intelligent harness could tailor per round message and budget. ideally you start with a small budget and expand based on some complexity criteria.

alas, i want to build things not related to AI.