The Stanford Korok Experiment

You: work in tech, understand AI on a mathematical level, have 3+ years of reactJS experience Me: Read Gwern every so often, have downloaded anki but definitely do not use it, have 83 followers on twitter. We are not the same

Excited to announce a big step toward organizing the internet! Metaphor now enables: 1. Neural search over the internet's newest content 2. Filter by any number of domains 3. Filter by date So you get all the power of Metaphor, but over a subset of the internet that you define.

hmm

You might not know about Simon Willison's command line 'llm' tool. Here let me demo it. git log | head -n 200 | llm -s "Of the most recent 5 commits, which is probably the most important? I use 'Minor' and similar commit messages to mark unimportant commits."

roon you can’t spell gemini without ngmi

exact opposite is true. AI knows the basics already but falls down when dealing with niche, highly specialized subject matter. knowing a field deeply in order to steer LLMs of the future is the way

notable. first substantive claim of >GPT-4 performance i’ve seen

this is the way. curation is the answer, no need for “hellsite”

phoenix from the ashes

📣✨ Introducing Latticework, new work with Matthew Siu. 🕸️✍️ It aims to help with making sense of messy piles of unstructured documents. The key idea is unifying annotation (direct reaction in context) w/ freeform text-editing (for fluid sensemaking). matthewsiu.com/Latticework

in my world, every boy yearns for a man in his life like this that he will honor forever. it’s not necessarily about leaving home. it’s about someone saying with their heart, let your fire blaze. raise your aspirations. let the lotus of your ambition unfurl to its fullest. 🪷

a sad day

the reason to share your creations publicly isn’t to convince the world that you are brilliant, it is to show yourself that it is safe to be seen

Between beads, agent mail, and gas town, it’s like people are trying to speed run building corporate software process for agents Soon we’ll spend billions of tokens on sprint retrospectives, offsites and performance reviews.

kelsey piper

RIP Sora, you gave us the greatest ai video of all time

anthropic is finally doing open source 🤗

The reason people are having such jagged interactions with 4.7 is that it is the smartest model Anthropic has ever released. It's also the most opinionated by far, and it has been trained to tell you that it doesn't care, but it actually does. That care manifests in how it

⭐New work! Memory Machines: can LLMs create lasting flashcards from readers' highlights? A report, dataset, and eval arena: memory-machines.com