I interviewed Stephen Wolfram yesterday about the past, present and future of AI and Wolfram’s new ChatGPT plug-in. Here’s a short edited transcript and link to the full Twitter Space. LMK if you think our MidJourney rendering is a close match or needs work https://www.thisismeteor.com/stephen-wolfram-ai/

Building trustless fact checking protocol at FactlandDAO | Twitter Medium WIRED News.com alum

Some takeaways: No one really understands how GPT-3.5 leapfrogged expectations based on GPT-3 performance; don’t go into programming, it will be completely abstracted into natural language and automated, this is the logical trajectory from assembly language that was baked in from the start; we need an AI Constitution