Tylor Steinberger

@triposat @joinmassive Is there any list of what websites you utilized to run this test available somewhere?

🚀 After an amazing time at Stellate, I've decided to start a new business. I've founded expand.ai, and we're in the current YC batch - S24! For techies: expand.ai instantly turns any website into a type-safe API you can rely on. 💻 For non-techies: We get you any dataset from the web. 📊

One of the best releases with the new set of HttpApi modules inspired by the early work on effect-http. Massive work first by @sukovanej followed by @tim_smart !

Create/update/delete local-first data 🗂️ and sync it to @neondatabase github.com/guillempuche/l… Tech stack: - @pnpmjs monorepo - @powersync_ (sync engine) - @kysely_ (query engine) + npmjs.com/package/effect… by @TylorS167 - @EffectTS_ (superman in @typescript). Repo w/ w/o Effect

@ecyrbedev @MichaelArnaldi I'd asked Tim previously about a sql-kysely package in the main repo and we weren't totally sure we'd want it there, but maybe that has changed. I'm more than happy to contribute anything there as well

@TylorS167 @MichaelArnaldi Oh nice. I'll take a look. I did not work on schema part yet. I was not sure if we needed it. My plan is to make a PR directly for Effect monorepo and not have it as a side community project.

@MichaelArnaldi i found a way to make both have the same syntax. 🥳 Thanks Javascript Proxy ! Left: effect/sql Right: Kysely Dialect

@localhost_5173 Shout out to @unscriptable for pushing the web forward!

The first article I can find about debouncing in JavaScript was published in 2009 by John Hann Websites before 2009:

Hey all, I'm officially open to new work opportunities! I'm looking for Senior and above TypeScript roles, ideally in some kind of technical leadership capacity. If you see anything that might be a good fit for me, feel free to share my name or send them my way. Thank you!

@steida @iamwil @schickling Koka has no such limitation. An "Effect" can seamlessly represent both types of abstractions.

@steida @iamwil @schickling Likely little to nothing in functionality terms, but Effect can not be used *directly* as the encoding for Streams. Streams in Effect are abstractions atop of Effect. Which is totally great, but interoperability does require conversion to and from Stream <-> Effect.

My 2 big (technological) bets for this decade: 1) Local-first: A different way how apps are architected – mainly how data flows and responsibility of client vs server. 2) Effect: A better foundation to build with TypeScript.

Effect is finally API Stable! Starting with 3.0 the main package will follow semantic versioning: → major releases will include breaking changes → minor releases will include new features and new modules → patch releases will include bug fixes

@RyanCarniato @_jessicasachs They came with the ES2015 standard. developer.chrome.com/blog/es2015-pr…

@_jessicasachs I have no idea how to find when that first shipped. Proxy is super loaded when you try to search for it.

Important dates in my Signals timeline so far: July 2010 - KnockoutJS Apr 2012 - Meteor Dec 2013 - S.js Feb 2014 - Vue Mar 2015 - MobObservable Nov 2016 - Svelte Mar 2017 - Surplus Apr 2018 - Solid Feb 2019 - React Hooks Sept 2020 - Vue Composition API

@iamwil @schickling Roughly speaking they're similar to single-shot effects found in something like multi-core OCaml, but not capable of multi-shot like something found in Koka.

@schickling Are these Algebraic Effects?

@jdegoes @GolemCloud Nice work! I'm wondering though, where are the data centers located?

I am delighted to announce the launch of @GolemCloud, a new cloud platform for deploying immortal, reactive, invincible workers. 🎉 Built on durable computing, Golem Cloud lets you create reliable, resilient, scalable cloud apps faster than ever before. golem.cloud/post/unveiling…

@MichaelArnaldi @TheEduardoRFS They are much slower than Effect. Namely in construction. But to answer your question, I switched to Effect because we were pointed in the same direction anyways and it's a lot easier to keep this stuff alive with the help of others.

@MichaelArnaldi @TheEduardoRFS Sorry, I actually meant to respond one message up in the thread. Regarding performance being unoptimized in the JS engines due to lack of usage.

Classic FP: principle of the least power Effectful FP: principle of the most power

@MichaelArnaldi @TheEduardoRFS Why do you think generators made it into JS if they're not expected to be used? They're initially there to power async/await. Same in rust too. It's surely pretty well optimized code paths. They're just expensive to create, and to run if they're in hot paths w/ sync code.

@TheEduardoRFS @TylorS167 did it on top of generators, he ended up switching to Effect later. Note that generators are mutable too, not easy to deal with them decently.