Unfalsifiable stories of peace | Mechanize War, Inc.
mechanizew.ar · 12,689 words · saved by 1 readers
Building RL environments to enable the full automation of armed conflict across the global economy of violence.
Mechanize War, Inc. Mechanize War, Inc. We build reinforcement learning environments that frontier AI labs use to train and evaluate their combat models. In these environments, models carry out military operations such as planning an invasion, coordinating a drone swarm, or negotiating a ceasefire they have no intention of honoring. An automated grader scores how well a model performed, and those scores become reward signals during training. Did you hold the city? Did you minimize casualties on your side? Did you file the after-action report before the deadline? All measurable. All gradable. O
saved by
related reading
- Android Dreamsandroid-dreams.ai
- What failure looks like — LessWronglesswrong.com
- Interactive: Anatomy Of An AI Kill Chainai-killchain.airwars.org
- The AI War and How to Win Italexw.substack.com
- The future of AI is already written | Mechanize, Inc.mechanize.work
- The Myth of the Human-in-the-Loop and the Reality of Cognitive Offloading - Perry World Houseperryworldhouse.upenn.edu
- Questions about the Future of AI - by Dwarkesh Pateldwarkesh.com
- The upcoming GPT-3 moment for RL | Mechanize, Inc.mechanize.work
- AI Risks that Could Lead to Catastrophe | CAISsafe.ai
- Mechanize, Inc.mechanize.work
- After Automation | Everyevery.to
- What will be left for us to work on?normaltech.ai