Presenter
Brendon Wong, Cosmic
I’m a social and tech entrepreneur passionate about societal transformation and life improvement. To advance societal transformation, I cofounded Roote, a $1M+ nonprofit working on improving societal systems. These systems range from mitigating the causes of existential threats to humanity (meta existential risk) to making digital feeds less addictive and divisive, and more informative and useful (Tweetscape). To advance life improvement, I founded Better, a social enterprise that researches, builds, and shares the best ways for people to improve their own lives and the world. I also run Antigravity Investments, an investment firm that increases funding for good causes by helping donors and charities better grow their savings...
Summary:
Brendon Wong discusses safety-first cognitive architectures and their relevance to security. These architectures are designed with AI safety in mind, ensuring interpretable and corrigible goals, plans, and actions. Different architectures, like Conjecture and Eric Drexler’s Open Agency Model, have varying levels of similarity to models like Auto GPT. Security implications include sandboxing and preventing influence between system components. Cognitive architectures restrict AI models’ access to necessary information and actions. Challenges and solutions depend on the design and underlying AI models. For Wong, as technology advances, it is important to reevaluate safety measures and develop new ones.
