100x defect tolerance: How we solved the yield problem
cerebras.ai
Why is Cloudflare Pages' bandwidth unlimited?
mattsayar.com
Ropey – A UTF8 text rope for manipulating and editing large text
github.com
I have made the decision to disband Hindenburg Research
hindenburgresearch.com
dnSpyEx: .NET debugger and assembly editor
github.com
Why the weak nuclear force is short range
profmattstrassler.com
Sweden brings more books and handwriting practice back to its schools (2023)
apnews.com
Lambda Calculus in 383 Bytes (2022)
justine.lol
Modern JavaScript for Django developers
saaspegasus.com
Code reviews: A success story
blogsystem5.substack.com
Empirical Health (YC S23) is hiring interns to build data-driven primary care
ycombinator.com
Google is making AI in Gmail and Docs free, but raising the price of Workspace
theverge.com
Show HN: I made a tool to save multimedia from various platforms
saveplays.com
Show HN: I built a fair alternative to Product Hunt for indie makers
Sutton Hoo and Syria: Anglo-Saxons who served in the Byzantine army?
academic.oup.com
Scheduled tasks in ChatGPT
help.openai.com
Sky-scanning complete for Gaia
esa.int
TikTok preparing for U.S. shut-off on Sunday
reuters.com
Israel, Hamas reach ceasefire deal to end 15 months of war in Gaza
reuters.com
US will ban cancer-linked Red Dye No. 3 in cereal and other foods
bloomberg.com
Train faster static embedding models with sentence transformers
huggingface.co
Very interesting. I wonder if they tried this on classification as well. They mention but then don't talk about it in detail.