Katja reads to you
Katja Grace reads other people's stuff out loud. This is not meant to be public. If you aren't friends with Katja, it probably isn't for you.
Episodes
12 episodes
Situational Awareness: 2. From AGI to Superintelligence: the Intelligence Explosion (Part B)
Situational Awareness: 2. From AGI to Superintelligence: the Intelligence Explosion (Part A)
AI progress won’t stop at human-level. Hundreds of millions of AGIs could automate AI research, compressing a decade of algorithmic progress (5+ OOMs) into ≤1 year. We would rapidly go from human-level to vastly superhuman AI systems. ...
Situational Awareness: 1. From GPT-4 to AGI: Counting the OOMs (Part H): The next four years/It's this decade or bust
Situational Awareness: 1. From GPT-4 to AGI: Counting the OOMs (Part G): The next four years
Situational Awareness: 1. From GPT-4 to AGI: Counting the OOMs (Part F): Counting the OOMs / Unhobbling
Situational Awareness: 1. From GPT-4 to AGI: Counting the OOMs (Part E): Counting the OOMs / Algorithmic Efficiencies
Algorithms are getting way better, but we're going to hit a data wall or jump over it
Situational Awareness: 1. From GPT-4 to AGI: Counting the OOMs (Part D): Counting the OOMs / Compute
Compute
Situational Awareness: 1. From GPT-4 to AGI: Counting the OOMs1 (Part C)
The trends in deep learning
Situational Awareness: 1. From GPT-4 to AGI: Counting the OOMs1 (Part B)
"GPT-2 to GPT-4"
Situational Awareness: 1. From GPT-4 to AGI: Counting the OOMs1 (Part A)
AGI by 2027 is strikingly plausible. GPT-2 to GPT-4 took us from ~preschooler to ~smart high-schooler abilities in 4 years. Tracing trendlines in compute (~0.5 orders of magnitude or OOMs/year), algorithmic efficiencies (~0.5 OOMs/year), and “u...