I really think everyone should read the GPT-4 whitepaper released by OpenAI. At a minimum: Pages 4-8: System capabilities Page 48: Chart showing you why guardrails are needed Pages 54-56: Emergent behavior Pages 58-59: Economic impact and acceleration

“Alignment” is often thought of as a future thing: to prevent ultra-smart AIs from taking over the world. But everyone really should read the white paper from Open AI (or check out this thread). I bet there is at least one thing unaligned AIs can do right now that will worry you

Source: [link] This comes up in the risk section of the GPT-4 paper. Malevolent uses of this technology are scary. But man, if you're using it ethically...it's so powerful.