Board

I offer a high-level model for why software work feels like it sucks now: we don't treat software like creative output.

Editor’s note: This is the second of a two-part article. It will mostly make sense even without reading Part 1, but you would be missing out on a lot of context! Before we start, let's take a quick detour to talk about that word in the heading below. Japanese

Editor’s note: Due to the sheer amount of interesting material here we decided to break this into a two-part article. All across Japan, one sight remains ubiquitous – the convenience store. A far, far cry from their American cousins, convenience stores in Japan are without exception, spotless, well-stocked, open 24x7,

In the fast-moving open-source world, programs can come and go quickly; a tool that has many users today can easily be eclipsed by something better next week. Even in this environment, though, some programs endure for a long time. As an example, consider the PostgreSQL database system, which traces its history back to 1986. Making fundamental changes to a large code base with that much history is never an easy task. As fundamental changes go, moving PostgreSQL away from its process-oriented model is not a small one, but it is one that the project is considering seriously.

A bewildered audience is better than a bored one.

It's rarely the terrible decisions, processes, or even people that'll sink your organization. It's the accumulation and inertia of the mediocre ones. Dealing with the truly bad is easy. It's painfully obvious to all that change is required. The danger is imminent. It's much harder to find the will to act when the danger lurks in inadeq...

A CI workflow starts when a developer pushes a proposed change (pull request, changeset, patch, etc.). The code goes through a cycle of reviews and testing until it passes automated and manual (i.e., review) tests. Then it gets merged into the main branch.

More startups die of indigestion than starvation — Bill Hewlett In tennis, unforced errors are shots that should have been easy to return but were missed nonetheless. They aren’t attributed to an opponent's strength but are the result of our mistakes. Unforced errors are subtle mistakes that compound over time.

A vertical operating system possesses the same qualities as a system of record, but with a few key additions that make it a “premium” system of record.

Since it took us years to get into the cloud in the first place, I originally imagined it would take us years to get out as well. But all that work to containerize our applications and prepare them for the cloud actually turned out to make it relatively easy to exit. And now, after six months of effort, it's done. We're out. The last a...

First movers rarely win markets in technology. Ramp is a real competitor to Brex. DoorDash was founded years after GrubHub, Seamless, and Postmates. We use Excel instead of Lotus (or Visicalc). Facebook beat Myspace (and Instagram "beat" Facebook). When Google was founded in 1998, there were at least 20 other venture funded search engines. Google wasn't even the first to do pay-per-click advertising on its search engine either (that was Goto.com). Why is fast following so effective? Fast follo

Through several recent discussions of GUI programming in Rust,I have been left with the impression that the term ‘GUI’ means significantlydifferent things to...

How many humans does it take to make tech seem human? Millions.

I really like the "tech debt" metaphor. A lot of people don't, but I think that's because they either don't extend the metaphor far enough,...

Lots of mistakes, some uptime too.

Sequence diagrams really shine when you’re documenting different parts of a system and the various ways these parts interact with each other.

OpenAI and Microsoft have been crowned as the winners of the generative AI wave. Google is seen as having made severe missteps. Pundits say that incumbents will accrue all the value. There are no moats for startups. But ChatGPT is only 7 months old. Incumbent AI features are working, but we don’t know the counterfactual. New models are being trained and deployed. Open-source models aren’t as good but are catching up or have exceeded proprietary ones in different dimensions (developer tooling, ecosystem, etc.).

Abilities unused will atrophy, so putting in anything less than your best means giving up on what you’re capable of. You can’t save talent or energy for better days, only watch it go to waste. This is a hard truth to accept if you don’t think your company has earned your best. And companies make employees feel like that all the time. T...

Image diffusion models like Stable Diffusion can generate all types of images — from photorealistic quality to stylized images (e.g., in the style of Studio Ghibli or Leonardo Da Vinci). We started to control the output of these models with prompt engineering — finding the perfect combination of tokens to mix and match to achieve great images. Then we started curating the dataset — e.g., fine-tuning with images taken with a DSLR camera. Low-rank adaptation (LoRA).

What would a token-efficient language for LLMs look like? YAML is pervasive as a configuration language partly because of its human readability over JSON. But — what’s more token-efficient if you’re using an LLM? Let’s look at some configuration languages as an example, but the concept also generalizes to programming languages.

Gamedev.net is the leading resource for game developers, featuring daily news updates, over 1500 featured articles and tutorials, and the most active game development forums anywhere!

The main thing is to keep the main thing the main thing — Stephen Covey An important lesson I’ve learned over the years (and still refining). A few dimensions in my experience. The first is prioritization — actually picking the main thing. It’s often hard at first; there are so many competing priorities. And usually, the initial answer is horizontal scaling rather than vertical scaling — i.e., hire more, allocate more resources, and try more things. But putting some work into prioritizing. Do one thing, and do it well. Ruthlessly cut everything else.

Let me share a little theory I have about optimism, and why progress is so easy to underestimate.

The cloud leader confirmed that two new services introduced Tuesday at re:Invent can be used to manage applications on Microsoft and Google's cloud services.

Learn about Expensify's unusual story and how we evolved to create the best money management app in the world. Learn who we are and much more at Expensify today!

From corporate BeReal to an iMessage read count, Soren Iverson is coming up with totally unhinged updates for your favorite apps.

Despite an exponential increase in computing power, our applications require more machines than ever because of architectural decisions made 25 years ago. You can eliminate much of your complexity and cost by using SQLite & Litestream for your production applications.