We analyzed >80 LLMs in the deep dive blog post from DevQualityEval v0.6 for generating quality code. Check out the insights and results 👇

Geometry reveals all. I’m super proud of our new paper. Let me share some highlights: 1/ Language models develop “Dyck heads” that learn to do nested bracket-matching (which the linguists call "Dyck languages") x.com/georgeyw_/stat…

Starting in 2010, Google has signed 115 20-year power purchase agreements for wind and solar project with a total capacity of 14 GWe. "However, Google has since recognized that the approach is limited. 'PPAs are often isolated from broader grid planning and utility investment

Colibri has reached 100 meters! 🚀🎉 In 60 seconds, it climbed to 105m, diverted 30m north, and safely landed back to its pad. This is the flight we promised from day one, while no reusable rocket has flown freely in Europe yet! We did it with a tiny team and under 250kCHF! 😉

This is what I've been saying! Modern tech design seriously undervalues low latency. They make architectural decisions that make a responsive experience very difficult to achieve, and end up building pretty but annoyingly laggy crap

Isn't this a pretty big bearish signal for standard RLHF for capabilities

Ongekend. Volgens Zuid-Koreaanse geheime dienst gaan er 12,000 Noord-Koreaanse soldaten naar Oekraine, waaronder speciale eenheden. Geruchten hierover gingen al rond, maar er komt nu rap veel bewijs naar boven.

I was looking up some history behind the x86-64 transition, particularly around the Pentium 4 time frame, and I found out that Bob Colwell (Pentium Pro chief architect) has been posting on Quora. Pentium 4 had a version of x86-64 that was fused off.

using github.com/kolbytn/mindcr…, we added Claude 3.5 Sonnet and Opus to a minecraft server. Opus was a harmless goofball who often forgot to do anything in the game because of getting carried away roleplaying in chat. Sonnet, on the other hand, had no chill. The moment it was

Turns out Don Pettit may celebrate his 70th birthday IN SPACE next April, as NASA’s oldest active astronaut. Pretty sure that’s a first.

incredible how much people underestimate what a carnage road traffic is.

I hate ads that portray unvirtuous behavior in a favorable light Apple has run a couple ads recently where the central narrative is "our technology will help you get away with lying" I keep seeing these on TV, they're really toxic imo

Google has a nice 'Dark web report' similar to Have I Been Pwned at myactivity.google.com/dark-web-repor… Also linked from the security tab of your Google Account webpage.

Good narrative on attackers leveraging time-pressure to get you to do things you normally would not.

A story I like telling is how I ran the AV at a place and when I deployed adblock the detections fell off a cliff. I wish more people had this visceral "holy shit" moment. "It was this easy the whole time?" ~2012. I recently deployed uBlock Origin to an F500, on my insistence.

Well yes. All Plus users in the EU, Switzerland, Iceland, Norway, and Liechtenstein now have access to Advanced Voice.

Beyond computer use, the new Claude 3.5 Sonnet delivers significant gains in coding—an area where it already led the field. Sonnet scores higher on SWE-bench Verified than all available models—including reasoning models like OpenAI o1-preview and specialized agentic systems.

Fun story from our time working on computer use: We held an engineering bug bash to make sure we found all the potential problems with the API. This meant bringing a handful of engineers in a room together for a few hours. We were hungry so one of our engineers' first computer

The EngineAI SE01 can only walk straight ahead?

Why is the internet so slow right now?