GPT-3
May 2020 Gwern.net newsletter: GPT-3 scaling, implications, deep theory; anime GAN updates, and 1 book review.
Posts that commenters keep linking to, ranked by how many different people have cited them in Hacker News comments. 6,985 posts.
May 2020 Gwern.net newsletter: GPT-3 scaling, implications, deep theory; anime GAN updates, and 1 book review.
I have three big announcements to make today. The first is the official announcement of our next project. We’ve been quietly talking about it over the past few months, but today we want to tell you a bit more about it and finally reveal its name:
Why can’t we just use vector-based icons in our application? This question, in a variety of reinterpretations, keeps on resurfacing in various forums, blogs and articles. It is asked in the context of desktop applications where the same icon is used in different ways, starting from very small 16*16 icons in file…
I won’t discuss anything in this post. I’m tired of discussing technical things with people with skewed opinions and monetary interest. I’ve talked enough in the Bitcointalk forum…
We are past the event horizon; the takeoff has started. Humanity is close to building digital superintelligence, and at least so far it’s much less weird than it seems like it should be. Robots...
One of my most popular posts was also my very first… which has since been lost to the Great Blog Devouring of ’14. So… here’s the post from over ten years ago (has it been that long?!?!?!) about an event from almost twenty years ago. Still stuff I have to remind myself about some days.
Analyzing predictors of Google abandoning products; predicting future shutdowns
For my first post on dev.to I thought I'd write about a nice, safe topic that's free of controversy:...
smallpt is a global illumination renderer. It is 99 lines of C++, is open source, and renders the above scene using unbiased Monte Carlo path tracing (click for full size) .
Letters between John W Backus and Edsger W Dijkstra, 1979
A comprehensive introduction to investing in stocks, mutual funds (managed and index), short-selling, and stock options.
In 2001, Edsger W. Dijkstra wrote a letter to the Budget Council of The University of Texas. A PDF is available here , I’ve typed it up so that everyone can read it. Sadly, the curriculum was changed to Java . Relatedly, the algorithmic language Scheme was replaced by Python in MIT’s The Structure and Interpretation…
Here, I provide an in-depth analysis of GPUs for deep learning/machine learning and explain what is the best GPU for your use-case and budget.
Now that Sparrow is effectively dead many of its users will be looking for a new email client. If you're not afraid of the terminal you may want to give Mutt a try.
Which values of the Agile Manifesto do you find foolish? If you're going to tar and feather it(Agile), maybe you should start at the source.
He worked with computers; she worked with trees, and the flowers that took hold on the sides of the Mountain.
X-Windows is the Iran-Contra of graphical user interfaces: a tragedy of political compromises, entangled alliances, marketing hype, and just plain greed. X-Windows is to memory as Ronald Reagan was to money. Years of "Voodoo Ergonomics" have resulted in an unprecedented memory deficit of gargantuan proportions…
JDK 16 is out, and as usual, each new release comes with a bunch of new features, enhancements and bug fixes. ZGC received 46 enhancements and 25 bug fixes. Here I’ll cover a few of the more interesting enhancements.
Yesterday, Sam Altman and Greg Brockman were fired from the Board of Directors of OpenAI.
Yesterday I submitted a post on Hacker News out of frustration at moderators pedantically changing post titles. After quickly rising to the number two spot on the front page with 482 points and 37 comments, the post was deleted by a moderator, presumably the operator of the site, Paul Graham.
Integer overflow detection/handling in Rust is sometimes misunderstood.
In the previous post, we discovered the potential for data loss in RabbitMQ clusters . In this oft-requested installation of the Jepsen series, we’ll look at etcd : a new contender in the CP coordination service arena. We’ll also discuss Consul ’s findings with Jepsen.
Do you love ("very") high-level languages? Like Lisp, Smalltalk, Python, Ruby? Or maybe Haskell, ML? I love high-level languages.