These two papers seem to be sniffing in the right direction. arxiv.org/html/2604.04... arxiv.org/html/2602.12...
How Well Do Agentic Skills Work in the Wild: Benchmarking LLM Skill Usage in Realistic Settingsarxiv.orgПрофиль
yonicd
Профиль VivelyI tend to use Johnny 5 as my LLM written narration
So all the corporate cybersecurity ppl who manage firewalls can all just go to lunch… their services are not needed any longer
God bless his dumb dumb … so dumb… heart slate.com/technology/2...
The people want Hassett to break out the cubic model!
Wall street wanted less regulation…. Stoves be touched
That is very apparent with users main complaint for llm generated slides. The token vs utility ratio is insanely unbalanced. Futzing with slide layouts and bad prose vs messaging and effective story telling. Reminds me back in the day stake holders focusing on palletes, shapes and size of plots
Are we all just Barney at this point?