Lessons from the hacks
The article discusses the risks associated with rapid AI development, emphasizing the need for greater transparency and collaboration to balance innovation with safety. It highlights that technology companies prioritize growth and scalability, increasing risks, while governments may act slowly or overreact once harm occurs. Frontier models, despite being open, may exhibit unpredictable behavior due to increased inference-time scaling and assumptions about user intent. The industry is unprepared for the next 12-24 months, underscoring the importance of open models and societal preparedness to manage AI risks.
More from Interconnects
I wrote an AI textbook — how long until AI can do it better?
Reflections on AI's writing ability and how AI models get more capable.
interconnects.ai··Mads Kristian Nylund·AI EducationAI GovernanceAI Ethics
5 useful things you'll learn in my new post-training textbook (shipping now!)
After a few long years of finding time to document my lessons from training open models, my post-training book is done!
interconnects.ai··Mads Kristian Nylund·AI EducationAI DevelopmentAI Tools
Introducing our Artifacts Hub and Adoption Dashboard
Scaling our curation and measurement of the open ecosystem.
interconnects.ai··Mads Kristian Nylund·Open Source AIAI ToolsAI Infrastructure