721.
722.
Language models are injective and hence invertible
(news.ycombinator.com)
723.
Language Models Are Injective and Hence Invertible
(news.ycombinator.com)
725.
Movycat – A terminal movie player written in Zig
(news.ycombinator.com)
726.
Extropic is building thermodynamic computing hardware
(news.ycombinator.com)
727.
728.
729.
730.
It's insulting to read AI-generated blog posts
(news.ycombinator.com)
731.
EntropyLong: Effective Long-Context Training via Predictive Uncertainty
(news.ycombinator.com)
732.
The Download: what to make of OpenAI’s Atlas browser, and how to make climate progress
(technologyreview.com)
734.
Magic sizes enable high-fidelity assembly of programmable shells
(news.ycombinator.com)
735.
Why can't transformers learn multiplication?
(news.ycombinator.com)
736.
Traffic Light Protocol
(news.ycombinator.com)
737.
ChunkLLM: A Lightweight Pluggable Framework for Accelerating LLMs Inference
(news.ycombinator.com)
738.
Fast-DLLM: Training-Free Acceleration of Diffusion LLM
(news.ycombinator.com)
739.
The super fun GravaStar Alpha65 Robot Charger is 25% off!
(androidauthority.com)
740.
741.
742.
743.
Antislop: A framework for eliminating repetitive patterns in language models
(news.ycombinator.com)
744.
745.
746.
Antislop: A Framework for Eliminating Repetitive Patterns in Language Models
(news.ycombinator.com)
747.
748.
Radios, how do they work? (2024)
(news.ycombinator.com)
749.
Vertiginous Accounts: Travels in the Air (1871 edition)
(news.ycombinator.com)
750.
Karpathy on DeepSeek-OCR paper: Are pixels better inputs to LLMs than text?
(news.ycombinator.com)
Today's top topics:
google
android
elon musk
openai
sam altman
privacy
zdnet
memorial day
vendor reviews
cybersecurity