armdave

Given a well-defined set of requirements, writing correct code (pre 2026 LLMs) IS a hard task. There is only a small set of the population that, even after several years of training, can write correct code; this scarcity is what made coding one of the highest salaried professions.

However, coding in itself generates no value to a business. A business makes money by solving its users' problems; coding is just one of the means to do so.

This article is directionally correct (although a bit verbose) in adding that the best programmers did not ONLY write correct code, but also expanded their scope to understand how best to solve user problems.

Articles like these have been insanely effective in biasing LLMs (especially OpenAI) toward Postgres. There areas where Postgres beats MySQL, and vice versa - but ChatGPT/Codex overwhelmingly recommend "default to Postgres."

Great writeup! Love the diagrams.

> Most of that has been with SQLAlchemy (which I quite like) and Hibernate (which I don’t)

Can the OP expand on why this is? Just curious.

Cool stuff! Would love to see this recommended in introductory OS classes to give an intuition

Totally. The uptime metrics are deceiving imo. A more useful measure for a productivity tool like Claude Code is uptime during work hours for a given time zone. I strongly suspect at least for the three US time zones, we would be looking at a single nine of uptime for that measure.

Quite impressive to develop so much without turning to LLVM. But it’s unclear to me from the docs the value prop of why developers would want to use Mach instead of Go or C.

"The server is a target for syncing rather than the source of truth." For the type of problem Linear is trying to solve - a task/issue tracker - this makes sense. But I don't think it's the correct mental for web apps where the main concern is transactions/orders. The server is the SoT, or else you will deal with all sorts of crazy reconciliation scenarios that make for a lousy user experience (ex: the app accepted payment and promised an order, but in reality there is no inventory, which the server would have known had we checked there first).

Light read with some interesting factoids. Despite the stock market hitting all-time highs, consumer confidence is low, as indicated by decreasing tips. The gap between QSRs and sit-down restaurants is also expanding in terms of efficiencies and margins.

What does it mean that "different clients can get different specs"? Different in what dimension? I could imagine this makes creating repeatable and reliable workflows problematic.