Leo Gorodinski

@eulerfx

Founder @CohesiveSystems, Prev: Co-Founder / CTO alvys.com @AlvysTeam, Engineer @JetTechnology

Del Mar, CA

cohesive.sh

Joined May 2008

Leo Gorodinski
@eulerfx
15h
Perhaps what’s changing isn’t the value of reuse, but what we should reuse. We’ve historically optimized for reusable realizations: libraries, components, implementations. If realization becomes cheap, that matters less. But reusable semantics…the invariants, constraints, and
Jaana Dogan ヤナドガン
@rakyll
17h
No one truly talks about the many foundational values in software engineering was about reusability. If reusability doesn’t matter that much, it impacts every fundamental value we’ve been holding for decades.
Leo Gorodinski
@eulerfx
Aug 22
Perhaps when the dust settles, we’ll concede that English alone doesn’t scale as the “programming language” for intent. But until we measure, it’s all conjecture. I especially like that PostHog is measuring accuracy, and whether the semantic layer compounds rather than becoming
PostHog
@posthog
Aug 21
Article
Building a semantic layer: What it is and how we did it at PostHog
Ask Claude, Cursor, and PostHog AI the same question: "what was our MRR last month?", and you'll get three different queries and three different numbers. I did exactly this, half-expecting the tools...
Leo Gorodinski
@eulerfx
Aug 21
There are two different concerns here. One is stochasticity: an LLM may produce different outputs for the same prompt. The other is underdetermination: consequential decisions aren’t specified, so the model makes them for you.
Ryan Peterman
@ryanlpeterman
Aug 20
Why Microsoft didn't have LLMs do their recent Typescript native rewrite in Go Anders Hejlsberg (Creator of Typescript): "If we just let AI loose on, what is it, half a million lines of code that we have in the old compiler? Well, I don't know that that would absolve us from
00:00
Leo Gorodinski
@eulerfx
Aug 20
I think this is true, but there’s a graveyard of sophisticated libraries that were technically right and still lost on adoption. The answer isn’t weaker foundations…we need them to build increasingly complex systems. The answer is a simpler surface: map technical concepts to
Michael Arnaldi
@MichaelArnaldi
Aug 20
Effect is verbose is an argument that makes no sense, it is verbose compared to code that only cares about the happy path, that is not testable and doesn’t integrate telemetry. Compared to production grade code Effect is terse.
Leo Gorodinski
@eulerfx
Aug 19
I wonder how much of this changes as AI writing becomes ubiquitous. If what matters most is the idea, will we eventually care less about whether a human wrote every sentence, much as we may care less about whether a human wrote every line of code? English is different because
Paul Graham
@paulg
Aug 19
I was thinking about how I recognize AI writing, and one big tell is excessively colorful verbs. A handful of journalists might write that a proposal "drew" 100 votes, but any normal person will just say "got".