This research makes the claim that using “Chain of Thought” prompting for LLMs does not generalize. I consider this evidence of LLM limitations (at least for now?). “This spotlights drawbacks of chain of thought, especially because of the sharp tradeoff […]
“To dramatically oversimplify the internet, every product on the Internet is just one of two things…” I didn’t realize it at the start of my career, but one reason lots of other businesses didn’t make sense to me is because […]
“For the analysis, I used an LLM to embed the companies’ value propositions in a vector space, ran a clustering algorithm on top of the embeddings, manually analyzed the results, and used ChatGPT to edit the draft. The conclusions are […]
“Each of these cycles has been larger and lasted longer than the last, and I want to be clear: each cycle has produced genuinely useful technology. … There is an initial burst of rapid improvement, followed by gradual improvement, followed […]
Did you see that Apple ad? The one where they crushed all the creative tools of the last several centuries into one super-thin iPad. Creatives didn’t love it. This Stratechery piece uses the ad as a jumping off point for […]
Yes! Well, almost! Okay, soon! Fine, someway! “People in the industry tend to think that flight is useful and awesome, and not necessarily in that order. One of the reasons that the idea of flying cars has endured is that […]
Google’s DeepMind research team made something really good. You might enjoy flipping through this paper and reading the exhibits. “Finally, Med-Gemini’s performance suggests real-world utility by surpassing human experts on tasks such as medical text summarization and referral letter generation, […]
“This sequence of events illustrates how the AI arms race is making geopolitical divides simultaneously more fraught and simpler. In its quest to defend US companies’ IP and prevent China from accessing the most advanced capabilities, the US government is […]
Human forecasting accuracy gets much better when aggregating predictions from a lot of individuals. LLMs are especially bad forecasters compared to these human crowds. But what about crowds of LLMs? “Our finding opens the door for simple, practically applicable steps […]
He was an early investor in Twitter and Tumblr, and always seems to be a bit ahead of the rest of us on digital publishing innovation. He migrated his long-running personal blog to the blockchain and isn’t looking back. ~ […]