Commercial LLM Agents Are Already Vulnerable to Simple yet Dangerous Attacksarxiv.org·1 pts·TaurenHunter·0
"Creating a Government That Works Better and Costs Less" by VP Al Gore [pdf] (1993)cia.gov·17 pts·TaurenHunter·6
Bolt: Bootstrap long chain-of-thought in LLMs without distillation [pdf]arxiv.org·15 pts·TaurenHunter·5
Test-time scaling new approach: extra test-time compute improves LLM reasoningarxiv.org·2 pts·TaurenHunter·0
DeepSeek spent more than it claims: 1.6B in nearly 100k GPUs/chipstaiwannews.com.tw·7 pts·TaurenHunter·5
Could/should Jubilee debt cancellations be reintroduced today? [2019 pdf]econstor.eu·2 pts·TaurenHunter·0
DeepSeek R1-Zero's self-verification and search abilities in just 3B parametersnitter.lucabased.xyz·3 pts·TaurenHunter·0
The Greatest Growth Hack of All Times Meets Its David in a Chinese Quantcentreforaileadership.org·2 pts·TaurenHunter·0
Russia is producing in 3 months what the whole of NATO is producing in the yearnato.int·7 pts·TaurenHunter·1
Chicago immigrants stay home from work to avoid potential ICE arrestschicagotribune.com·7 pts·TaurenHunter·0
Enhancing Chat Language Models: Scaling High-Quality Instructional Conversationsarxiv.org·1 pts·TaurenHunter·0
The art of illusion: the role of decoys in military deception and modern warfare [pdf]finabel.org·1 pts·TaurenHunter·0
Universal-2-TF: All-Neural Text Formatting for Automatic Speech Recognitionarxiv.org·1 pts·TaurenHunter·0
Nuance inhibits abstractions and is not a virtue of good sociological theory [pdf]kieranhealy.org·2 pts·TaurenHunter·0