Mercury ran a profile of me last week, written by Brooke Jackson-Glidden, which is generous and gets the thing I care about right. The part worth extending is the mechanism. Every Graze feed gets built by somebody who knows more about its subject than a central ranking system ever
Almost two years ago I wrote a post here about a repo called sky-feeder, which at the time was about four days old and consisted mostly of a JSON schema and a great deal of optimism. The idea was that feeds ought to be declarative - that you should be
If the original sin of Web 1.0 was the pop-up ad, the original sin of web 2.0 was the move to algorithmic feeds. Opaque optimization strategies aimed at maximizing private revenue for the sake of what was otherwise externally billed as public goods became increasingly toxic, spawning
There's no shortage of analogies for explaining what an LLM is capable of - one of the best, though, is from this New Yorker article proclaiming it as a "blurry JPEG of the web". This metaphor is particularly useful for capturing many of the technical aspects of
Long short-term memory models (LSTMs) are a family of neural networks that are predominantly used to predict the next value given a historical chain of previous values. These can be numerical predictions (i.e. where is the stock price going based on historical stock data) or categorical predictions (i.
Recently, I gave a talk at the PIE/Autodesk space to help contextualize some thoughts that have been percolating with regards to the nascent introduction of API-based, widely available LLMs like ChatGPT. In the hype cycle, I've observed some pretty broad claims about what's happening
Over the summer, I got really interested in the Helium network. Unlike many other crypto-backed projects, there was something at least of articulable value being provided by the project. In short, the promise of the Helium network is in building a two-sided market. On one side, there are
Big news! My good friend Tim Hwang and I have started a company, International Persuasion Machines (IPM). Our company is based off several foundational principles: 1. The history of cybersecurity is, in essence, chasing threat actors up an escalating software abstraction pyramid (e.g. as we build new layers of
For the past year, I've been increasingly focusing on what I have come to call "sociotechnical security" - whereas "technical security" seeks to identify and remove unintended flaws in the architecture of platforms, "sociotechnical security" is all about identifying and removing the incentives
About 10 days ago, I saw a post for a ridiculously cute car of a make and model that I previously did not know existed: I knew about MGBs and how they had a fairly terrible reputation but this thing looked so slick. I started reading up on them and
@TAP_deals is a bot on Twitter that uses machine learning to extrapolate the estimated value of a plane based on historical trends from listings of all other planes on Trade-a-Plane.com. While the code underpinning the Twitter bot constantly evaluates all planes listed at Trade-a-Plane.
Large-Scale Missing Data in a Widely-Published Reddit Corpus As researchers study complex social behaviors at scale with large datasets, the validity of this computational social science depends on the integrity of the data. That’s why researchers have a duty to check the datasets rather than assume their