Browse Hacker News
6,826 matching stories
[ 146 ms ] page + count
- The meta-harness for coding agents (soloterm.com)
- Investment Seer: Agentic AI-Powered Stock and Crypto Analyst (investmentseer.com)
- Show HN: Hindsight: Automated Work Timeline (apps.apple.com)
Tired of trying to remember what you worked on yesterday for daily scrum? Me too. So I used a VLM to automate it — it quietly captures what you do and builds a timeline. Your choice of model, and with a local one…
-
Given the dangers of AI hacking (which must be directed) given that our oligarchs own the AI systems and are currently unregulated?
- HomeBody: A Humanoid That Explores, Remembers, and Acts on Its Own (tml.stanford.edu)
- Halo: Post-train LLMs 3x faster than TRL and Megatron (twitter.com)
- A Calendar of Smells (worldsensorium.com)
- What's Up with Liliputing? (liliputing.com)
- Classified Estimates Show the NSA Is Paying Billions to Test AI Models (washingtonsun.com)
- Hiring Smart (kk.org)
- Argus II: The story of a $150k artificial retina (youtube.com)
- Deep Scattering Layer (en.wikipedia.org)
- If Carpenters Were Hired Like Programmers (2013) (web.archive.org)
- Wiring memory into an agent costs tokens every turn. This one earned it back (vshulcz.github.io)
- Teardown: Herdr (github.com)
- The Tech Bro in the Treaty Room (politico.com)
- Accelerated Out of Core Shuffling (quasiben.github.io)
- Animated visual explanation of how LLM tokenization works (seeitflow.com)
- Fun: First-Class Functions, Currying, and a Surprise (blog.tinyinterpreters.dev)