A recent paper The Reversal Curse points out an apparent failure in large large language models like GPT-4. From the abstract: We expose a surprising failure of generalization in auto-regressive large language models (LLMs). If a model is
The Gemini protocol was a big part of my COVID-19 lockdown experience. Discovering this underground, small protocol, having long discussions on the mailing list, and most importantly for me, developing software. My terminal Gemini browser, Amfora, was my first public FOSS project, something actually intended for a…
In this long article, I, being as far as I know the first outside Microsoft to have learnt how the AARD code works, explain that I condemn its form but broadly sympathise with its function.
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.
🥇 Thanks to maya.land and goofpunk for sponsoring the blog! If you'd like to learn how to support my writing and get a shoutout here every week, check the sp...
I analyzed stackoverflow for secrets and leaks.
An introduction to the core ideas and approaches
GigaGPT is Cerebras’ implementation of Andrei Karpathy’s nanoGPT – the simplest and most compact code base to train and fine-tune GPT models.
An in-depth analysis that explains how this console works internally
In the ever-evolving field of information security, curiosity and continuous learning drive innovation. This blog series is tailored for those deeply engaged in experimental projects, leveraging Rust’s capabilities to push the boundaries of what’s possible. The focus on Rust, after exploring various programming…
Every person on the planet should have their own website, on their own domain name, and blog about whatever they want.
Google Search is dead to me. Here's how and why Kagi's high quality results and compelling value proposition allowed me to fully make the switch.
Check out the impressive results of the large-scale Folia test that took place on June 18th, 2023. Learn more about our findings and technical challenges in this post.
A detailed explanation of why Elixir uses a dot when invoking anonymous functions
Today Canonical is raising the bar for performance and observability by enabling frame pointers by default in Ubuntu 24.04 LTS. […]
50 Ohm impedance became the standard impedance used in RF transmission lines long ago, but it is still useful today and is a standard reference impedance used in test equipment.
GPT4-V is the new mode of GPT-4 that allows you to upload images as part of your conversations. It’s absolutely brilliant. It also provides a whole new set of vectors …
On Saturday 11th March I wrote about how Large language models are having their Stable Diffusion moment. Today is Monday. Let’s look at what’s happened in the past three days. …
This post is the perfect starting point for learning to build your own LTE network. View this lession as a guided introduction – including the installation, configuration, and best practices that will ease your learning.
William Gropp, University of Illinois; Pavan Balaji, Meta; Rajeev Thakur, Yanfei Guo, Kenneth Raffenetti, and Hui Zhou (all of Argonne National Laboratory), receive the ACM Software System Award for MPICH, which has powered 30 years of progress in computational science and engineering by providing scalable, robust…
1. I mostly don’t play chess anymore — it’s too addictive, and tends to take over your brain in a way I don’t like — but one habit I’ve retained is solving puzzles.
We’re excited to announce a new feature that allows users to seamlessly purchase and manage domains directly through Bluesky.
Roman soldiers weren't paid in salt, and they weren't given "salt money". Dictionaries of the 1700s and early 1800s are to blame for this myth.
Microsoft has made tons of noise about their Microsoft 365 Copilot toolset. They’ve made a lot of promises about its capabilities to enhance individual and team productivity, but access to it…
Someone once asked Napoleon how he decided where to assign soldiers. Napoleon's reply was that it's simple: soldiers are either smart or dumb, lazy or energetic. The smart and energetic I make field commanders. They know what to do and can rally the troops to do it. The smart and lazy I make generals. They
Google has been in the spotlight for the past few weeks with reports that it has been testing removal news links from search results. The move sparked outrage from MPs, who grilled executives earlier today at Canadian Heritage committee. But now it appears Google has company: the Globe and Mail reports that Facebook…
