The Accumulation of Regulation
Like many who do business in California, I often complain about the regulatory burden (free at last!) People will ask, "So what one regulation would you get
DeepSeek Didn’t Invent the Wheel — It Just Put a New Sticker on It
AI has a hype problem. Every few months, a new model drops, and people start asking if it’s the next big thing. OpenAI, Anthropic, Google…
Analytical Intelligence Is Coming - Developer forever
The release of Deepseek R1 has made headlines, but the unsung hero is its sibling: R1 Zero. While R1 showcases reasoning through human feedback, R1 Zero demonstrates something far more transformative—reasoning without human feedback. This works only in analyti…
UV tricks
Don't forget to uv self update before trying those
To Hit DeepSeek 0
Update: I was using the following models: deepseek-r1:7b and llama3:latest (7b) DeepSeek-R1 came out generally over the weekend, and a lot was talked about the cost to generate such a complete and capable model. I am happy to see that it is an open model tha…
Building My E-Ink Tea Labeling System - Tynan.com
My life sometimes seems to consist of long periods of intense effort to help me save seconds or minutes from my day. This latest project must be the most extreme example of that. I store my tea in identical metal tins. Years ago I went down a rabbit hole of de…
Allegory of the Tsunami
Change is on the horizon, AGI approaches
Give People Something To Disagree With - Don't Break Prod
Don't Break Prod provides bite-sized pieces of career advice for software engineers.
Polymarket Paradox | Rakshith Aloori
Take this poll.
Spice up with Nix: Scripts with magical dependencies | yukiisbo.red
Hey there, Instead of reading this outdated article, I highly recommend consulting nix.dev’s First Steps which is official documentation! F…
Manager vs Craftsman — Derrick Persson
Programming in the era of LLM's
Why I'm Ditching poetry for uv
How I cut my Python working environment's headaches in half with Astral's uv - a Rust-powered Swiss Army knife for Python workflows.
GPT1 Implemented with NumPy : Generate Jokes
I love implementing DL algorithms with NumPy, it’s best way to learn the engine. I’ve implemented SHAP, and CNNs, and GradCAM methods with pure numpy before, you can refer to my Github Page for this.
Thinking in community systems
Community can be a bit fluffy. A bit intangible. A bit directionless at times. We can believe in something, yet actioning it into reality is a whole other story. This year I've learned we can do better. We can be more focused. We can double down on what matte…
Why (Open Source) DeepSeek's shock is important
How a Chinese startup is challenging AI giants with open-source innovation, achieving ChatGPT-level performance at 2% of the cost. Is this the future of accessible AI?
Making LLM workflows human friendly
Validate the data as it streams, don’t make your users wait.
What we got wrong about the cloud
Not everyone bought my comparison of how the cloud developed and AI is developing that I posted the other day . But I wanted to flesh out s...
Building a (T1D) smartwatch from scratch
Documenting an attempt to build a single-purpose smartwatch.
Audio Waveforms in Kdenlive: Technical Upgrades for Speed, Precision, and better UX
I recently worked on audio waveform generation in Kdenlive, an actively developed free and open-source non-linear video editor. It boasts a user base of over 100,000 users per release on Linux, Windows, and macOS. The Kdenlive video editor. To improve user exp…
Writing a .NET Garbage Collector in C# - Part 1
First part of a series of articles about writing a .NET Garbage Collector in C# using NativeAOT. This part sets the expectations and setups the project, dealing with the first difficulties.
You Shall Not Get Access 🧙🏻♂️: Browser Permissions
Web Security Educational Blog
The Content Overflow Era – the end of the Long Tail?
What follows might be trivial by now, but it is always good to put a word on it. I'm speaking about...
Is Colorado's EV incentive worth it?
I wasn't even looking for a new car when my coworker told me, 'You can lease a Nissan Leaf for like $20 a month.' That sounded too good to be true. Spoiler alert: it was. How are electric vehicle leases so cheap here in Colorado?…
Why Gambling is (more) Rational (than you think)
I remember watching a video a while ago where a pro card counter answers questions about gambling. In one of his answers, he said that what makes someone a true gambler is that the joy of winning a bet is more than the sadness from losing a bet. This has stuck…
Semantic Complexity Gap
Another place for thought infusion
Eric Chiang | Confidential Compute and GPUs
Recently, I’ve had a few conversations about NVIDIA Confidential Compute, usually in the context of startups trying to sell security products to AI companies.
correcting my model
go home
Seven things I know after 25 years of development
The transcript of the talk given at EuRuKo'24
SLAP and FLOP
The SLAP and FLOP Address and Value Prediction Attacks
Use SQL Server .NET Framework CLR objects with SDK based cross platform .dacpac builds
SQL CLR objects are assemblies that are created in .NET Framework languages and deployed to the SQL Server. The SQL CLR objects can be used to extend the functionality of the SQL Server by creating stored procedures, functions, triggers, and user-defined types…
Downtown Doug Brown » The invalid 68030 instruction that accidentally allowed the Mac Classic II to successfully boot up
This is the story of how Apple made a mistake in the ROM of the Macintosh Classic II that probably should have prevented it from booting, but instead, miraculously, its Motorola MC68030 CPU accidentally prevented a crash and saved the day by executing an undef…
ggml : x2 speed for WASM by optimizing SIMD
PR by Xuan-Son Nguyen for `llama.cpp`: > This PR provides a big jump in speed for WASM by leveraging SIMD instructions for `qX_K_q8_K` and `qX_0_q8_0` dot product functions. > > …
Open Source with C#
Turn your videos into live streams with https://restream.io Let's work on SharpSite, an open source CMS that you can contribute to and learn from. https://sharpsite.org
Run DeepSeek-R1 Dynamic 1.58-bit
DeepSeek R-1 is the most powerful open-source reasoning model that performs on par with OpenAI's o1 model. Run the 1.58-bit Dynamic GGUF version by Unsloth.
The CORRECT Way to Handle Value Objects in EF Core
Want to master Clean Architecture? Go here: https://bit.ly/3PupkOJ Want to unlock Modular Monoliths? Go here: https://bit.ly/3SXlzSt Get the source code here: https://www.patreon.com/milanjovanovic In this video, we dive deep into a common but frustrating iss…
Is Hexagonal Architecture Good for a Simple App?
💌 Free Developer Insights: https://guiferreira.me/newsletter Reddit: https://www.reddit.com/r/softwarearchitecture/comments/1hvebg2/should_a_simple_proxy_app_use_pure_ports_and/
Open-R1: a fully open reproduction of DeepSeek-R1
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
F: Graphics 3 (Category Compilation) #14～#20
In Category F: Graphics, I explain what goes into making art for games. Rather than sharing tool tutorials and the like, I offer practical points to consider when creating game visuals. [Index] 0:00 Opening 0:11 F-14 Motion Blur 2:39 F-15 Making the Imaginary…
Infrastructure Layer Should Always Be Logic-Free: Why?
Learn why your infrastructure layer should contain zero business logic and how to properly structure your .NET applications using Clean Architecture principles. Includes practical examples and common pitfalls to avoid.
The Alpha Myth: How Captive Wolves Led Us Astray
A wolf paces in captivity, teaching us all the wrong lessons about power
The NuGet packages we use to build elmah.io revisited
Explore the updated NuGet packages powering elmah.io! Discover new additions, changes, and essential tools driving our modern tech stack.
17-445 Machine Learning in Production
CMU course that covers how to build, deploy, assure, and maintain software products with machine-learned models. Includes the entire lifecycle from a prototype ML model to an entire system deployed in production. Covers also responsible AI (including safety, s…
The Illustrated DeepSeek-R1
A recipe for reasoning LLMs
The Taylorator - All Your Frequencies Are Belong to Us
For the past two weeks or so, I've been working on constructing the Taylorator. The Taylorator is a piece of software which allows me to flood the FM broadcast band with Taylor Swift's music. No matter where you tune your radio, you will only be able to listen…
Deepseek R1 Explained by a Retired Microsoft Engineer
Dave explains why Deepseek R1 is such a big deal, explains how it works, what's new, and brings you up to date on the implications and fall out! Free Sample of my Book on the Spectrum: https://amzn.to/3zBinWM Watch ShopTalk Now on Dave's Attic: https://youtu…
3 years of intense learning - The Dawnmaker Post-mortem
It's been 3 years since I started working on Dawnmaker full-time with Alexis. The creation of our first commercial game coincided with the creation of Arpentor Studio, our company.
Translation using deep neural networks (part 1) |
In this article, I’ll introduce language modeling using deep learning and will focus on the problem of translation.
Scaling Compute for RAG
Unlock higher RAG accuracy by strategically spending compute.
Computers should beep more
Sound can make you more productive if it's calm & played at the right moment. I saved 30min a week.
How we built Ellipsis (or: Lessons from 27 months building LLM coding agents) - Ellipsis.dev | Automated code reviews
Automated code reviews and bug fixes. Software engineers are more productive with Ellipsis, a developer tool that catches bugs, answers questions, and generates working, tested code.
Tim Cook Is Failing Us
I know this is a rather dramatic headline and that some people might incorrectly assume that I’m mad about some particular piece of hardware or software, but it should come as no surprise that I’m hopping mad at Tim Cook for how he’s ingratiating himself to Pr…
Scientific progress is severly understudied
Right now I’m interested in questions like Is scientific progress slowing down? and, if yes, what might be causing it?
WordPress as a Self-Hosting Platform – Alex Kirk
Leaving the current drama around WordPress.org aside for now, I believe that there is a use case for WordPress that is heavily underutilized: Using WordPress as a platform for self-hosting.
Champions League of Webdesign - Cutting-edge technology meets top-tier digital demands
This blog post highlights often-unseen details behind the high performance website of Escape Club Amsterdam, blending cutting-edge technology with design and commerce.
Non-random uniform disk sampling
Consider the problem of generating N points on a disk of diameter D. The function must be deterministic and the points somewhat uniformly spread around on th...
API Clients: Bruno betrays, Yaak yaks
There are a lot of options for a REST/HTTP/API Client. I believe Yaak is the best option today.
I Tried To Bootstrap A Startup - memoricide
Two years ago I started a business for quantified hiring for tech companies to hire engineers. Last Thursday I shut down the main server that runs candidate submissions. And nobody screamed.
What the F*** is a VAE?
Rehan Sheikh's personal website
My Afternoon Project Turned Into Four Days of AI Lies, USB Chaos, and Hard Lessons
A quick project spiraled into a marathon of AI missteps, hardware headaches, and hard-earned lessons. Welcome to the story of Deskthang.
New Years resolutions for PyTorch in 2025 : ezyang’s blog
In my previous two posts "Ways to use torch.compile" and "Ways to use torch.export", I often said that PyTorch would be good for a use case, but there might be some downsides. Some of the downsides are foundational and difficult to remove. But some... just s…
Vim sucks when you have ADHD
: Vim is fine, just not for me. Use the right tool that suits you best. For many years, I've been loyal to Vim. I hoped that with enough practice, I'...
InkLink - Nick Winans
Real-time Collaborative E-Paper Canvas
I Got Tired of My Own Inability to Focus on One Thing - Durmonski.com
You have to pick the places you don’t walk away from. But you also have to pick the places you do walk away from. Usually, the faster you do the latter, the better. My insatiable curiosity – about books, ideas, projects, etc. – is probably both my best and my …
On the Eve of Superintelligence
Preparing for AI Progress
The Rise of DeepSeek: What the Headlines Miss
TLDR: timing is political but tech is real, compute constraints bite differently than you think, and the story is more complex than "export controls failed."
Kosay Jabre
Why do restrooms stalls in North America have such huge gaps in the door?