Alignment whack-a-mole: Finetuning activates recall of copyrighted books in LLMs
(news.ycombinator.com)
1141.
1142.
Finetuning Activates Verbatim Recall of Copyrighted Books in LLMs
(news.ycombinator.com)
1143.
A grounded conceptual model for ownership types in Rust
(news.ycombinator.com)
1144.
A Grounded Conceptual Model for Ownership Types in Rust
(news.ycombinator.com)
1145.
White House Opposes Anthropic’s Plan to Expand Access to Mythos Model
(feeds.content.dowjones.io)
1146.
1147.
1148.
Vera: a programming language designed for machines to write
(news.ycombinator.com)
1149.
1150.
1151.
Official SAP npm packages compromised to steal credentials
(bleepingcomputer.com)
1152.
Opus 4.7 knows the real Kelsey
(news.ycombinator.com)
1153.
1154.
Mistral Medium 3.5
(news.ycombinator.com)
1155.
It runs Doom: AI chatbot edition
(engadget.com)
1156.
Why friendly AI chatbots might be less trustworthy
(feeds.bbci.co.uk)
1157.
The friendlier the AI chatbot the more inaccurate it is, study suggests
(feeds.bbci.co.uk)
1158.
1159.
Compositing and Blending – Exploring the math and intuition behind blend modes
(news.ycombinator.com)
1160.
Show HN: Rip.so – a graveyard for dead internet things
(news.ycombinator.com)
1161.
1162.
1163.
We decreased our LLM costs with Opus
(news.ycombinator.com)
1164.
1165.
Friendlier LLMs tell users what they want to hear — even when it is wrong
(feeds.nature.com)
1166.
Pervasive and programmed nucleosome distortion on single chromatin fibres
(feeds.nature.com)
1167.
Uncertain dynamic response of mid-latitude winter precipitation
(feeds.nature.com)
1168.
1169.
How to build custom reasoning agents with a fraction of the compute
(venturebeat.com)
1170.
OpenAI models coming to Amazon Bedrock: Interview with OpenAI and AWS CEOs
(news.ycombinator.com)
Today's top topics:
google
android authority
anthropic
apple
meta
openai
android
clickfix
microsoft
android 17