What Do We Want to Read Next?
Choose Our Own Adventure
Hi it’s me again.
I’ve been traveling and haven’t had the focused time to weave together another long post. Despite the delay, this slower pace may help with the core intention of this newsletter—I want to be incredibly thoughtful and dig into frameworks, analysis, and implications rather than share quick newsy items about AI. Still, this newsletter will only thrive if I post so I’m happy to say that I’ll be back home in just a few days. I’ll be returning home with plenty of notes and drafts swirling around from my travels to San Francisco and company retreat.
Now I want to ask you, what’s next? We’ve got exciting topics in the pipeline. Here are a few:
Framework for building with AI: This one will be about how to assess and think about AI tools. We’ll explore using a concrete example how we might go from prototype to robust product featuring prompt optimization via reinforcement learning from human feedback (RLHF). This all may sound fancy but it can all be prototyped in Google sheets and is intended for a broad audience. You’ll choose your own adventure based on how technical you can go.
LLMs lowering the cost of thinking: In this article we’ll be discussing how LLMs will lower perhaps the costliest compute on Earth, human thinking. On a personal blog post, I’ve already made the point that white-collar work (especially from developed countries) has the biggest carbon costs and should be optimized to get the most value out of knowledge workers of this world.
Path to AGI and beyond: Here I’ll explore what the path to achieving artificial general intelligence and artificial superintelligence may look like. It will be at a bird’s eye view (so no knowledge of ML papers needed) but there will be explorations into how it might all come together. I’ll examine the two emerging schools of thought of an integrated vs. modular approach to developing AGI.
Defence against the rouge AI: We get a little philosophical in this one and compare humans to LLMs and organizations (e.g. corporations) to AGI. I’ll pose a few philosophical thoughts about how we might want to steer AGI development to avoid crisis.
Hope you all have a good week!