Don't tell me RAG is easy Retrieval-augmented generation (RAG) is helpful for AI workflows, but it certainly is not easy.
q value · Rohit Gupta Understanding p value can be confusing. People who get it have an edge over those who don’t.
The discreet charm of the infrastructureless Early in the morning in a European city, I awoke and wondered for a moment where I was. The room was totally black apart from a small glow f...
Microsoft Dependency Has Risks This is not an opinion, this is not an opinion... it's an opinion, isn't it?
An All-Around Better Horse by Patrick Hebron An illustrated essay that reimagines design and engineering processes and software through the transformative power of AI.
Running large language models at home with Ollama - Chris Beckman Deploy state-of-the-art LLMs locally with Ollama! This guide walks through how to set it up, then dives into practical workflows: offline embedding search, …
Evaluating Long-Context Question & Answer Systems Evaluation metrics, how to build eval datasets, eval methodology, and a review of several benchmarks.
The Nature of Lisp, Code Generation and Wieldable Programming Power The Nature of Lisp article and thoughts on programming language power vs usability and code generation.
From NixOS to Ubuntu This year, I moved from NixOS to Ubuntu. This post outlines my experience through that process, as well as a new philosophy for how I manage my machine and configuration which embraces the defaults and more contemporary applications.
What Problems to Solve - By Richard Feynman A former student, who was also once a student of Tomonaga’s, wrote to extend his congratulations. Feynman responded, asking Mr. Mano what he was now doing. The response: “studying the Coherence theory with some applications to the propagation of electroma...
Getting an LLM to Set Its Own Temperature I tried to get an LLM to pick a random number, but it was bad at it. So I invented a whole new sampling technique to fix this.
Nuenki | The more LLMs think, the worse they translate Testing LLMs across multiple different thinking strategies demonstrated a consistent regression as they thought more, with the exception of Ensemble strategies.
4 Month Journey to CISSP - 2025 So, I haven't posted much in the past four months. That’s because my head has been buried in a study guide while working toward ISC2's CISSP. I've completed quite a few certifications in the past—starting with entry-level ones and moving into some offensi...
Why, in 2025, do we still need a 3rd party app to write a REST API with Django? This is a question the predident of the DSF was asked at FOSDEM after his talk. Let's try (very hard) to NOT answer it because... Do we really need a 3rd party app to write an API with Django?
Understanding SVG Coordinate Systems and Transformations (Part 1) — The viewport, viewBox, and preserveAspectRatio — The personal website of Sara Soueidan, inclusive Web UI engineer
Rules Clobber Goals - kupajo One workout doesn’t make you an athlete any more than one cigarette makes you a smoker. But you are an accumulation of your recurring choices and actions. We set goals to motivate ourselves. But, of course, we’re better at setting goals than reaching them...
Divine Provenance To our good fortune, Aristotle did not write, "It is the mark of an educated mind to be able to entertain a thought without accepting it," because this quotation, by sociologist Lowell L. Bennion from a mistranslation of Aristotle, offers many lessons. It...
Daniel's Blog · Encouraging the use of LLMs made interviews easier (for us as interviewers) The advent of cluely and the ever growing arms race between the interviewer and the interviewee is a fascinating phenomenon to me. Betting that people would prefer cheating or that companies needed screen for cheaters because the use of LLMs is heavily di...
How to find a job in tech in 2025? · Vibe Interview Prepare for job interviews with Vibe Interview's AI-powered tools. Generate tailored cover letters, practice with recruiter call simulations, and track applications—all in one platform.
Public bet Motivation I’m a very insecure guy. In fact, that was first real feedback that I received from a manager (now a dear friend), “You’re exceptionally good in your tasks, but damn, you’re too insecure about yourself”, this trait got in my way in a variou...
So You Want to Learn LLMs? Here's the Roadmap: A Real-World, No-Bloat Guide to Building, Training, and Shipping LLMs · Osman's Odyssey: Byte & Build The straight-up, no-BS roadmap for learning LLMs in 2025. Skip the ML fluff and endless prerequisites. Get the actionable phases, projects, and resources to actually build, train, and ship large language models—from the ground up.
OpenAI Charges by the Minute, So Make the Minutes Shorter I discovered a fun and strangely obvious trick for summarizing videos faster and reducing costs: just speed them up. Cheaper, faster OpenAI transcriptions with a little ffmpeg trick.
Gemini CLI: your open-source AI agent Free and open source, Gemini CLI brings Gemini directly into developers’ terminals — with unmatched access for individuals.
Thick Coins by Theodore Nichols Hello this is Theodore Nichols and welcome to my world wide web page.
GitHub - microsoft/edit: We all edit. We all edit. Contribute to microsoft/edit development by creating an account on GitHub.
ASP.NET Core in .NET 10 Preview 4: JSON Patch with System.Text.Json .NET, ASP.NET Core, Visual Studio, Azure, Jaliya Udagedara
The National Archives at College Park, Maryland We hold permanent records created by Federal agencies that include: Textual records from civilian agencies Army unit records dating from WW1 Navy unit records dating from WW2 Still pictures Electronic records Cartographic and architectural holdings Motio...
Fun with uv and PEP 723 How to use uv and the Python inline script metadata proposal PEP 723 to run scripts seamlessly.