Saved by Simon Joliveau Breney
Fwd: AI Alignment Is Censorship
🚨 AI Policy Alert: The German Federal Office for Information Security publishes the report "Generative AI Models - Opportunities and Risks for Industry and Authorities." Quotes & comments:
"LLMs are trained based on huge text corpora. The origin of these texts and their quality are generally not fully verified due to... See more
The solution is a personalized counter-intelligence LLMs that each person owns and controls.
Exogenous neural content filtering model; artificial cogsec aug.
Of course this is impossible if a handful of monocultural companies control all models.
Beff – e/accx.com
I had a talk with my big boss ML engineer at work about the new deepseek RL model that dropped and how we can use a reasoning model in our agentic projects.
He sent me an arxiv paper and said we aren’t allowed to use tech from threat nations(China) as security in LLMs is a new field and we can’t take risks at F500. I... See more
A big brain dump of hastily interconnected thoughts on the coming AI revolution…
1) The Web is for AI
People ask: “What happens on a world where the cost of writing goes down?”
The answer: “The cost of human reading goes up.”
For... See more
Zach Tratarx.com