I want everything local – Building my offline AI workspace
instavm.io
Ultrathin business card runs a fluid simulation
github.com
Tor: How a military project became a lifeline for privacy
thereader.mitpress.mit.edu
Build durable workflows with Postgres
dbos.dev
GPU-Rich Labs Have Won: What's Left for the Rest of Us Is Distillation
inference.net
Astronomy Photographer of the Year 2025 shortlist
rmg.co.uk
Json2dir: a JSON-to-directory converter, a fast alternative to home-manager
github.com
Ask HN: How can ChatGPT serve 700M users when I can't run one GPT-4 locally?
How we replaced Elasticsearch and MongoDB with Rust and RocksDB
radar.com
Apple's history is hiding in a Mac font
spacebar.news
HRT's Python fork: Leveraging PEP 690 for faster imports
hudsonrivertrading.com
Apollo 13 Commander Jim Lovell has passed away
nasa.gov
Job growth has slowed sharply; the question is why
stayathomemacro.substack.com
Getting good results from Claude code
dzombak.com
Linear sent me down a local-first rabbit hole
bytemash.net
The surprise deprecation of GPT-4o for ChatGPT consumers
simonwillison.net
Fire hazard of WHY2025 badge due to 18650 Li-Ion cells
wiki.why2025.org
Over engineering my homelab so I don't pay cloud providers
ergaster.org
A robust, open-source framework for Spiking Neural Networks on low-end FPGAs
arxiv.org
Open SWE: An open-source asynchronous coding agent
blog.langchain.com
How Attention Sinks Keep Language Models Stable
hanlab.mit.edu
You need view or gesture pointers to go make "go from here to there" meaningful . A standalone sound interface will always fail .
And for 3d space it needs ai, as omly ai can deduce the depth along the view axis by analysing context.
Then you need a cubic volume selection . With a relational addition .
"a,3,5 Select all except resting drones and fly to here."