When you're looking at student loan debt and an uncertain job market, while being asked to learn new skills, it's understandable to question what all that college was for in the first place. I'm not saying there's no value in a college education. Still, the co…
When crabs are thrown into a bucket, they instinctively grab at other crabs trying to climb out. As a group, crabs in a bucket pull each other down, preventing one another from escaping.
AndreGarzia.com website
Font Around and Find Out
two things LLMs can't understand
Dwarkesh Patel has recently interviewed Sholto Douglas and Trenton Bricken for a second time, and the podcast is very enlightening in terms of how the big AI labs think in terms of their economic strategy:
SQL is a declarative language; only the result of the query is specified. The exact steps to produce the result must be determined by the DBMS. Often, multiple ways exist to calculate a query result. For example, the DBMS can choose to use an index or perform …
My general advice for life is: be a good person, and care for the people around you. And follow this one very specific rule: avoid vendor lock-in.
Recently I’ve been working on a pretty big rust project and to my surprise Icouldn’t get tests to work properly.
Lately, I’ve been reflecting on the concept of being organic. To me, being organic means not forcing things to happen. It’s a mindset of alignment—understanding nature, respecting its pace, and all...
How we handle 92 million compilations a year without everything catching fire
At the time of writing (June 2025), the prevailing view in the software industry is that LLM-powered AI is either completely useless or will imminently destroy all software engineering jobs. As you might expect, the reality is somewhere in between. In this pos…
John Voorhees:
About a year ago, I rebuild my home server. It still runs on Kubernetes, but I moved away from traditional tooling associated with it. The goal was simplicity; and I made some opinionated choices to achieve it. For example, I deploy everything using Terraform …
A web-based JSON editor application
Get up to date on all of the .NET sessions from Microsoft Build 2025 covering .NET 10, C# 14, .NET Aspire, ASP.NET Core, Blazor, AI development, and more!
Personal blog of Chris Given, government technologist. These words are always written by a human; I have no one to blame but myself.
The story behind writing The BEAM Book—ten years, three publishers, and an open‑source sprint.
News update: Precious Plastic is in problems. You can decide the future of the project. Die or grow
If you’re a software developer, you’ve probably worked with diff files. Git diffs, Subversion diffs, CVS diffs.. Some kind of diff. You probably haven’t given it a second thought, really. You make some changes, run a command, a diff comes out. Maybe you hand i…
an AI researcher going back to school for immunology
https://www.youtube.com/watch?v=gxwyKJOLQRM
Hello, welcome to a new post! Today I want to talk about a very important topic that should never be missing in serious software development: unit tests. Unit tests are code we write to validate the logic of our program. These tests are characterized by eval…
29 May 2025
Poker and startups share more than late nights, caffeine, and the occasional bluff. Both worlds rely on decision making under uncertainty, calculated risks, and knowing when to fold and move on. So let's break down some poker jargon with a founder's twist.
User research will need to adapt, and fast. But those who adapt will have an unfair advantage. Up until now, a core value of research has been to de-risk software development. When building is expensive, mistakes are costly, especially when they’re caught late…
The LLM developer is a new specialized role, whose goal is to keep the complexity of creating and maintaining AI-based software solutions under control.
Continual learning is a huge bottleneck
Fabian Falck, ML researcher, Generative Models
I can't decide how to best express the equivalent of Haskell's Error monad and DO block in Ruby...
June 3rd 2025
You can use it to read, archive and organise newsletters. It's free, synchronised between your devices and open-source.
I argue that microlearning is not effective for long-term appreciation of knowledge.
I was training an engineering team at Apple in 2010 after a minor, but important, technology advancement - and they were pissed.
A rant
Imagine for a moment that we paused all economic growth. The pie isn’t getting bigger, but it isn’t getting smaller either. Now, imagine we invent an efficiency improvement—some new technology. This means the same amount of goods and services can be produced…
As you know, this blog is more focused on sharing code snippets than on teaching, so today I’m going to show you something I recently discovered.
Founder and developer.
At Weights, we handle millions of inference and training requests every day. Users frequently use and train LoRa models, resulting in nearly a petabyte of user-generated models.