As AI models extend their capabilities to solve more sophisticated challenges, a new scaling law known as test-time scaling or inference-time scaling is emerging. Also known as AI reasoning or long-thinking, this technique improves model performance by allocating additional computational resources…

Intelligence, ±artificial AI versus the brain and the race for general intelligence We already have an example of general intelligence, and it doesn't look like AI. John Timmer – Mar 3, 2025 7:00 AM | 79 Story text Size Width * Links * Subscribers only Learn more There's no question that AI systems…

The Barn Swallow’s Flight Whenever you feel that there is no joy in the world, you must go to a place with swallows. Few things are more euphoric than a swallow in flight. The sun flashes on their iridescent backs. They swoop and dive and wheel around in reckless arcs. Up close they fly nearly too…

This article is part of our exclusive IEEE Journal Watch series in partnership with IEEE Xplore. Swarms of autonomous robots are increasingly being tested and deployed in complex missions, yet a certain level of human oversight during these missions is still required. Which means a major question…

News March 03, 2025 From the American Bar Association The ABA rejects efforts to undermine the courts and the legal profession Share: Three weeks ago, the American Bar Association spoke to you about values that guide us. We called upon every lawyer to insist that the government adhere to four major…

DeepSeek has made a lot of noise lately. Their R1 model, released in January 2025, outperformed competitors like OpenAI’s O1 at launch. But what truly set it apart was its highly efficient infrastructure—dramatically reducing costs while maintaining top-tier performance. Now, they're coming for data…

In late fall, I was in a break room with 10 or so autoworkers at the Avon Lake Ford plant near Lake Erie. As they talked about the challenges they and their families faced, one worker asked everyone whom they were voting for in the presidential election. One by one, they went around the table: the…

Welcome back to the weekly bonus content for “One First.” I usually put Thursday’s bonus content behind a paywall as an added incentive for those who are willing and able to support the work that goes into putting this newsletter together every week. But for reasons that will become obvious, I…

Contents Whitepapers January 1, 2024 Observability for Large Language Models - Understanding & Improving Your Use of LLMs Observability for Large Language Models shows you how to build a production feedback loop into your LLM development cycle to accelerate refinements and make your product…

Marriage is hardly a new invention. But leaving your husband is, relatively speaking. There’s one political party in particular that isn’t so happy with that development. During the 2024 election, vice presidential candidate J.D. Vance called no-fault divorce “one of the great tricks that I think…

This month we have a mailbag question from a reader who asks: Hi James, Your newsletter "being in the details" resonated with me. I would like to ask your opinion about managers writing code. I skim PRs but don’t critically review them and don’t write code. Should EMs write code in their day job?…

Michael Froman is president of the Council on Foreign Relations. Technological supremacy is a core pillar of American power. Conceivably, part of the “America First” agenda should be keeping America first when it comes to maintaining that edge. We owe much of our unmatched geopolitical and economic…

For years, the conventional wisdom for wide-eyed youngsters about to enter the job market was "learn to code." Now, it seems like some of the programmers themselves could use the same advice. That's according to Namanyay Goel, an experienced developer who's not too impressed by the new generation of…

At first glance, Frank Herbert’s Dune (1965) might appear to be a mere copy of the story of Lawrence of Arabia with some science-fictional window dressing. Several critics have pointed to the similarities between Lawrence and Paul Atreides—both are foreign figures who immerse themselves in a desert…

Giles' blog Posted on 24 February 2025 in Blogkeeping My post about blogging as writing the tutorial that you wished you'd found really took off on Hacker News. There were a lot of excellent comments, but one thing kept coming up: what's the point in blogging if people are using ChatGPT, Claude and…

gwern3mo*Ω4011836 Idea for LLM support for writing LessWrong posts: virtual comments. Back in August I discussed with Rafe & Oliver a bit about how to integrate LLMs into LW2 in ways which aren't awful and which encourage improvement---particularly using the new 'prompt caching' feature. To…

A few weeks ago I wrote about Git’s help.autocorrect setting and the strange tale of the origin of it’s deciseconds value. It got me to thinking about other git config settings that most people likely don’t know about and which should probably be defaulted differently. In this post, I’ll go through…

Did dogs domesticate themselves? (Image credit: martinagebarovska via Getty Images) Scientists don't know exactly how wolves were domesticated into early dogs, but it's possible that they domesticated themselves by choosing to coexist with humans so that, a new study finds, they could be fed…

There are two claims I’d like to make: LLMs can be used effectively1 for listwise document ranking. Some complex problems can (surprisingly) be solved by transforming them into document ranking problems. I’ve primarily explored both of these claims in the context of using patch diffing to locate…

(This document is the result of a series of discussions, some online and some in person, held between Robert "Uncle Bob" Martin and John Ousterhout between September 2024 and February 2025. If you would like to comment on anything in this discussion, we recommend that you do so on the Google group…

The AI treated the chip design as one complete system rather than a collection of parts. (Image credit: Princeton University) Engineering researchers have demonstrated that artificial intelligence (AI) can design complex wireless chips in hours, a feat that would have taken humans weeks to complete.…

With thousands of apps and billions of downloads, Flathub has a responsibility to help ensure the safety of our millions of active users. We take this responsibility very seriously with a layered, in-depth approach including sandboxing, permissions, transparency, policy, human review, automation,…

👋 here's an excerpt from the Manning book I'm writing that I think you'll like. In his famous Things You Should Never Do essay, Joel Spolsky says never to attempt a rewrite. It cannot succeed and you're throwing away lessons learned. Lessons you'll need to re-learn again. But research and experience…

What concerns me about Sacca's rhetoric is that he claims this is happening "right now:" I think other than the highest level of computer science, almost all of the rest of coding is fucking useless. You and I can go to ChatGPT and be like, 'I wanna build an app that does this, this, and this, and…

Over the past few months, thousands of builders have created websites and web apps using tools like Bolt, Lovable, and v0. The results, shared across X, Reddit, and YouTube, seem nothing short of magic: people are building interactive, dynamic products — and even generating revenue – without coding.…

Text embeddings, particularly modern embeddings generated from large language models, are one of the most useful applications coming from the generative AI boom. Embeddings are a list of numbers which represent an object: in the case of text embeddings, they can represent words, sentences, and full…

Thomas Frank, historian, journalist, and author of What’s the Matter with Kansas? and Listen, Liberal, joins us to dissect how Democrats abandoned populism, the rise of Trump’s faux-populism, and why the party refuses to embrace the working class. He also explores the path forward for authentic…

This is about a 17 minute read. Feel free to pause and come back to it later. Clojure is not one of the handful of "big" mainstream languages. This means that sometimes people are surprised that we are all in on Clojure. Why go against the grain? Why make it harder for yourself by building on niche…

INTERVIEW Inside Sam Altman’s ChatGPT HQ, they’re changing the world — again The next AI transformation will be driven by ‘agents’ that can book tickets, draft emails or conduct PhD-level research, say Open AI’s boss and his inner circle ILLUSTRATION BY TONY BELL Danny Fortson Sunday February 16…

There are times when Washington still feels like what it used to be—a sleepy little Southern city. Generations of some of the most powerful people on earth have spent their days in DC griping that the city is a backwater. But Washington’s modest position in our constellation of great cities was…

84 Shares We are witnessing an unprecedented pile-up of unicorn startups that have not raised any money since 2021. Currently, an estimated 517 global unicorns — or private companies valued at $1 billion or more 1 — raised their last known round more than three years ago, per Crunchbase data. Such…

I figured out a really simple pattern for experimenting with new Python packages today: uvx --with llm --with sqlite-utils ipython This command will work if you have uv installed and nothing else. It uses uvx to install and run the excellent ipython Python REPL in a new, dedicated virtual…

The global dollar system has few national winners. In the typical frame, understanding the US dollar means understanding the “exorbitant privilege” it confers on America. But the role of the dollar in structuring the international financial system and defining the relationship between a hegemonic US…