Peter Adams

Effective memory, as they point out, is created to the degree a model is able to predict the future. That’s the paradox of memory. The focus therefore isn’t restructuring memory contents - but having a stronger model with higher reasoning. No need to overcomplicate. Just let an intelligent model be aware that transcripts are available and something like grep exists and it will do perfectly fine. A bonus would be a strong vector index for your transcripts. You’re not missing out on a lot of you don’t have a «dreaming pipeline».

Companies not giving their employees unlimited AI usage - capping their radically increased productivity.

Reading books gives you a ridiculously unfair advantage. But it goes further. Of the small proportion of adults who read books, most merely consume them. Writing doesn’t just give you an unfair advantage. It rigs the game. It is the best way to organize your thoughts, practice articulating them convincingly, and sharpen your thinking. You stop passively consuming books and start wrestling with every sentence, paragraph, and chapter, approaching them with the penetrating sharpness and lucidity of a critical mind. With AI, that unfair advantage only multiplies. It can retrieve information related to what you’re reading to fuel your thinking. You can load the contents of a book into its context and turn it into an intellectual sparring partner. Given this unfair advantage, why don’t more people do it? From my experience, thinking—actually thinking—is painful. Because once you start, you’re confronted with your own ignorance. Your sense of identity is challenged and may even collapse. The ideal is to become an intellectual honey badger: one that ruthlessly attacks everything it encounters by default, questions it, exposes its flaws, and, most importantly, refuses to limit its diet. If you want to reinforce your sense of identity, read the books your family buys you for Christmas and the ones the algorithm tells you you’ll enjoy. If you want to be dangerous, read books that raise your pulse just by looking at them. Books that make you uncomfortable. Books your priest wants to burn. Books that would earn you ugly stares from your friends. Books you disagree with.

I'm fascinated by my lack of shock and awe to the announcement of a flying car. Why? It's what's happened the last month or two. AGY with Astra. Figure 3's doing the dishes. Neuralink turning thoughts into words on a screen. AI swarms solving important math problems. They put a fruit fly brain in a simulated world. First trial proving the possibility to reverse aging in humans. I think I'm just coming to terms with the fact we're entering a sci-fi world at this point.

I would trust the AI more than a human «expert» acting alone. Having said that I’d prefer a human expert using AI. Also, the definition of «understand» become extremely murky if we really think about it. Based on the common definition, I’d say we don’t understand fully most of the mechanisms of medicine.