Dev Wells
@wlls.dev
(╯°□°)╯︵ ┻━┻
Principal Eng @ Judi Health -- focused on frontend
Principal Eng @ Judi Health -- focused on frontend
I need to stay off of LinkedIn y'all. The amount of borderline web development hype-driven disinformation is getting ridiculous.
I'm going to lose it if I see one more AI-generated comparison image of how the React `use` API CHANGES DATA FETCHING FOREVERRRR
(no disrespect to the use API)
I'm going to lose it if I see one more AI-generated comparison image of how the React `use` API CHANGES DATA FETCHING FOREVERRRR
(no disrespect to the use API)
December 20, 2025 at 7:45 PM
I need to stay off of LinkedIn y'all. The amount of borderline web development hype-driven disinformation is getting ridiculous.
I'm going to lose it if I see one more AI-generated comparison image of how the React `use` API CHANGES DATA FETCHING FOREVERRRR
(no disrespect to the use API)
I'm going to lose it if I see one more AI-generated comparison image of how the React `use` API CHANGES DATA FETCHING FOREVERRRR
(no disrespect to the use API)
Reposted by Dev Wells
Testing server-rendered HTML:
Make a request, look at the result.
Testing React et al:
Make a request.
Receive HTML.
Ignore it!
It also contains a string.
The string is attached to a bird
The bird speaks latin
Find an antiquities professor
He can distract the bird
While you build a fake browser
Make a request, look at the result.
Testing React et al:
Make a request.
Receive HTML.
Ignore it!
It also contains a string.
The string is attached to a bird
The bird speaks latin
Find an antiquities professor
He can distract the bird
While you build a fake browser
December 19, 2025 at 8:59 PM
Testing server-rendered HTML:
Make a request, look at the result.
Testing React et al:
Make a request.
Receive HTML.
Ignore it!
It also contains a string.
The string is attached to a bird
The bird speaks latin
Find an antiquities professor
He can distract the bird
While you build a fake browser
Make a request, look at the result.
Testing React et al:
Make a request.
Receive HTML.
Ignore it!
It also contains a string.
The string is attached to a bird
The bird speaks latin
Find an antiquities professor
He can distract the bird
While you build a fake browser
Does AI-assisted coding actually make you faster, or does it just feel that way? Tired of anecdotes, time to experiment.
I built a CLI to find out. Define hypotheses, work in time-boxed blocks, track both feelings and git metrics.
Now running my own N=1 experiment.
wlls.dev/blog/devex
I built a CLI to find out. Define hypotheses, work in time-boxed blocks, track both feelings and git metrics.
Now running my own N=1 experiment.
wlls.dev/blog/devex
wlls.dev
Software engineer, tinkerer, writer.
wlls.dev
December 19, 2025 at 10:20 PM
Does AI-assisted coding actually make you faster, or does it just feel that way? Tired of anecdotes, time to experiment.
I built a CLI to find out. Define hypotheses, work in time-boxed blocks, track both feelings and git metrics.
Now running my own N=1 experiment.
wlls.dev/blog/devex
I built a CLI to find out. Define hypotheses, work in time-boxed blocks, track both feelings and git metrics.
Now running my own N=1 experiment.
wlls.dev/blog/devex
December 18, 2025 at 7:01 PM
just had to talk to a chatbot to cancel a subscription.
the "cancel subscription" button (already hidden in some low contrast, inaccessible ellipses in a modal) opened a chat with a prompt for cancelling.
no other apparent way to cancel and had to talk to an LLM to get the button.
please, no.
the "cancel subscription" button (already hidden in some low contrast, inaccessible ellipses in a modal) opened a chat with a prompt for cancelling.
no other apparent way to cancel and had to talk to an LLM to get the button.
please, no.
December 18, 2025 at 10:53 AM
just had to talk to a chatbot to cancel a subscription.
the "cancel subscription" button (already hidden in some low contrast, inaccessible ellipses in a modal) opened a chat with a prompt for cancelling.
no other apparent way to cancel and had to talk to an LLM to get the button.
please, no.
the "cancel subscription" button (already hidden in some low contrast, inaccessible ellipses in a modal) opened a chat with a prompt for cancelling.
no other apparent way to cancel and had to talk to an LLM to get the button.
please, no.
This year I built a little holiday celebration app for my team in SvelteKit and Gleam.
It was a blast! Gleam is really neat, highly recommend. I gotta try Lustre still.
Read about it here: wlls.dev/blog/snowglobe
Or try the (redacted / open sourced) version here: github.com/devdumpling/...
It was a blast! Gleam is really neat, highly recommend. I gotta try Lustre still.
Read about it here: wlls.dev/blog/snowglobe
Or try the (redacted / open sourced) version here: github.com/devdumpling/...
wlls.dev
Software engineer, tinkerer, writer.
wlls.dev
December 17, 2025 at 9:07 PM
This year I built a little holiday celebration app for my team in SvelteKit and Gleam.
It was a blast! Gleam is really neat, highly recommend. I gotta try Lustre still.
Read about it here: wlls.dev/blog/snowglobe
Or try the (redacted / open sourced) version here: github.com/devdumpling/...
It was a blast! Gleam is really neat, highly recommend. I gotta try Lustre still.
Read about it here: wlls.dev/blog/snowglobe
Or try the (redacted / open sourced) version here: github.com/devdumpling/...
Reposted by Dev Wells
Big improvement to frameworks on Cloudflare! You can now deploy most frameworks from the CLI with zero config. No adding adapters. No editing configs. No wrangler.toml! It's all done automatically. Experimental now, with 10 frameworks supported today.
developers.cloudflare.com/changelog/20...
developers.cloudflare.com/changelog/20...
December 16, 2025 at 3:50 PM
Big improvement to frameworks on Cloudflare! You can now deploy most frameworks from the CLI with zero config. No adding adapters. No editing configs. No wrangler.toml! It's all done automatically. Experimental now, with 10 frameworks supported today.
developers.cloudflare.com/changelog/20...
developers.cloudflare.com/changelog/20...
Reposted by Dev Wells
⚡️ The first Vite 8.0 beta is here!
- Powered by Rolldown, bringing significantly faster production builds and more consistency
- New features such as tsconfig paths or emitDecoratorMetadata support
- Bumping browser targets aligned with Baseline Widely Available
Read more in our announcement post!
- Powered by Rolldown, bringing significantly faster production builds and more consistency
- New features such as tsconfig paths or emitDecoratorMetadata support
- Bumping browser targets aligned with Baseline Widely Available
Read more in our announcement post!
Vite 8 Beta: The Rolldown-powered Vite
The first beta of Vite 8, powered by Rolldown, is now available. Vite 8 ships significantly faster production builds and unlocks future improvement possibilities. Read more about the migration, new fe...
voidzero.dev
December 3, 2025 at 12:18 PM
⚡️ The first Vite 8.0 beta is here!
- Powered by Rolldown, bringing significantly faster production builds and more consistency
- New features such as tsconfig paths or emitDecoratorMetadata support
- Bumping browser targets aligned with Baseline Widely Available
Read more in our announcement post!
- Powered by Rolldown, bringing significantly faster production builds and more consistency
- New features such as tsconfig paths or emitDecoratorMetadata support
- Bumping browser targets aligned with Baseline Widely Available
Read more in our announcement post!
Reposted by Dev Wells
Oxfmt: Oxc Formatter Alpha
A collection of high-performance JavaScript tools written in Rust
oxc.rs
December 1, 2025 at 3:43 PM
Has anyone tried actually running experiments on how they develop with LLM-assisted workflows? I see many subjective takes (pro and negative) but think it'd be neat to record a significant stretch of:
- no AI usage
- "moderate" usage (e.g. autocomplete, lite chat)
- heavy usage (IDE + agentic)
- no AI usage
- "moderate" usage (e.g. autocomplete, lite chat)
- heavy usage (IDE + agentic)
November 29, 2025 at 5:45 PM
Has anyone tried actually running experiments on how they develop with LLM-assisted workflows? I see many subjective takes (pro and negative) but think it'd be neat to record a significant stretch of:
- no AI usage
- "moderate" usage (e.g. autocomplete, lite chat)
- heavy usage (IDE + agentic)
- no AI usage
- "moderate" usage (e.g. autocomplete, lite chat)
- heavy usage (IDE + agentic)
Got asked at Thanksgiving by a family member how to get hired in tech. Like many juniors I've interviewed, their resume/github is all toy projects, which is good for learning but not hiring.
I tried to think about what I would do if I had no experience, but knew what I know now.
I tried to think about what I would do if I had no experience, but knew what I know now.
November 28, 2025 at 6:13 PM
Got asked at Thanksgiving by a family member how to get hired in tech. Like many juniors I've interviewed, their resume/github is all toy projects, which is good for learning but not hiring.
I tried to think about what I would do if I had no experience, but knew what I know now.
I tried to think about what I would do if I had no experience, but knew what I know now.
wlls.dev/blog/on-tail...
Every so often I get into a debate on tailwind (mostly yakshaving, IMO).
It goes back and forth, usually with people speaking past each other. I finally sat down and put my experience into words on why I like tailwind (and where I don't).
Every so often I get into a debate on tailwind (mostly yakshaving, IMO).
It goes back and forth, usually with people speaking past each other. I finally sat down and put my experience into words on why I like tailwind (and where I don't).
wlls.dev
Software engineer, tinkerer, writer.
wlls.dev
November 26, 2025 at 9:58 PM
wlls.dev/blog/on-tail...
Every so often I get into a debate on tailwind (mostly yakshaving, IMO).
It goes back and forth, usually with people speaking past each other. I finally sat down and put my experience into words on why I like tailwind (and where I don't).
Every so often I get into a debate on tailwind (mostly yakshaving, IMO).
It goes back and forth, usually with people speaking past each other. I finally sat down and put my experience into words on why I like tailwind (and where I don't).
Reposted by Dev Wells
Frontend is failing. 75% of devices with browsers are smartphones, but not even half of sites pass Core Web Vitals on them. Why not? Too much JavaScript, added to indulge SPA fantasies the data is falsifying in real time:
infrequently.org/2025/11/perf...
infrequently.org/2025/11/perf...
The Performance Inequality Gap, 2026 - Infrequently Noted
Embedded in this year's network and device estimates is hopeful news about the trajectory of devices and networks. It has never been easier to deliver pages quickly, but we are not collectively…
infrequently.org
November 25, 2025 at 5:23 PM
Frontend is failing. 75% of devices with browsers are smartphones, but not even half of sites pass Core Web Vitals on them. Why not? Too much JavaScript, added to indulge SPA fantasies the data is falsifying in real time:
infrequently.org/2025/11/perf...
infrequently.org/2025/11/perf...
Gleam seems like a much softer entrypoint into the BEAM than Elixir or Erlang for frontend engineers more used to Typescript/C-derived languages.
Has anyone made a component/JSX-like templating framework for it? Lustre is elm-y, which I like but is likely foreign to the average React dev.
Has anyone made a component/JSX-like templating framework for it? Lustre is elm-y, which I like but is likely foreign to the average React dev.
November 16, 2025 at 1:52 PM
Gleam seems like a much softer entrypoint into the BEAM than Elixir or Erlang for frontend engineers more used to Typescript/C-derived languages.
Has anyone made a component/JSX-like templating framework for it? Lustre is elm-y, which I like but is likely foreign to the average React dev.
Has anyone made a component/JSX-like templating framework for it? Lustre is elm-y, which I like but is likely foreign to the average React dev.
Reposted by Dev Wells
A couple of weeks ago at @perfnow.nl I gave a talk called "Fine-grained everything". It's about @svelte.dev (obviously), and the new frontiers of framework-land, but it's also about some stubborn misconceptions in web performance circles
Gift link: conffab.com/presentation...
Gift link: conffab.com/presentation...
Fine-grained everything
<p>Modern frameworks like Svelte are fast thanks to signal‑based fine‑grained reactivity, but performance is more than 60fps. This talk covers new approaches to build fast, reliable, data‑efficient ap...
conffab.com
November 13, 2025 at 9:03 PM
A couple of weeks ago at @perfnow.nl I gave a talk called "Fine-grained everything". It's about @svelte.dev (obviously), and the new frontiers of framework-land, but it's also about some stubborn misconceptions in web performance circles
Gift link: conffab.com/presentation...
Gift link: conffab.com/presentation...
Some reflections on emergent technology discourse: wlls.dev/blog/rethink...
How can we have less polarizing discussions on new technologies/ideas in software? So often it seems we speak past each other, often oversimplifying problems in order to rep "our camp".
How can we have less polarizing discussions on new technologies/ideas in software? So often it seems we speak past each other, often oversimplifying problems in order to rep "our camp".
wlls.dev
Software engineer, tinkerer, writer.
wlls.dev
November 14, 2025 at 3:56 PM
Some reflections on emergent technology discourse: wlls.dev/blog/rethink...
How can we have less polarizing discussions on new technologies/ideas in software? So often it seems we speak past each other, often oversimplifying problems in order to rep "our camp".
How can we have less polarizing discussions on new technologies/ideas in software? So often it seems we speak past each other, often oversimplifying problems in order to rep "our camp".
Reposted by Dev Wells
👀 Navigation API will be available in all browsers soon 🎉
Completely redefines the APIs used to build client-side routers / SPAs
🫤 History API: popState, link click.preventDefault() ...
✅ Navigation API: clean, centralized way to intercept navigation events 👌
Completely redefines the APIs used to build client-side routers / SPAs
🫤 History API: popState, link click.preventDefault() ...
✅ Navigation API: clean, centralized way to intercept navigation events 👌
November 5, 2025 at 9:35 AM
👀 Navigation API will be available in all browsers soon 🎉
Completely redefines the APIs used to build client-side routers / SPAs
🫤 History API: popState, link click.preventDefault() ...
✅ Navigation API: clean, centralized way to intercept navigation events 👌
Completely redefines the APIs used to build client-side routers / SPAs
🫤 History API: popState, link click.preventDefault() ...
✅ Navigation API: clean, centralized way to intercept navigation events 👌
Reposted by Dev Wells
Reposted by Dev Wells
last chance to take the State of JS survey if you haven't already, it closes in a day or two
survey.devographics.com/en-US/survey...
survey.devographics.com/en-US/survey...
State of JavaScript 2025
Take the State of JavaScript survey
survey.devographics.com
November 3, 2025 at 2:59 PM
last chance to take the State of JS survey if you haven't already, it closes in a day or two
survey.devographics.com/en-US/survey...
survey.devographics.com/en-US/survey...
The State of JS 2025 survey is now open! survey.devographics.com/survey/state...
State of JavaScript 2025
Take the State of JavaScript survey
survey.devographics.com
October 27, 2025 at 2:11 PM
The State of JS 2025 survey is now open! survey.devographics.com/survey/state...
Reposted by Dev Wells
Someone explain Elixir / Phoenix LiveView to me. What do you like about it?
The model reminds me of sync engines, except Phoenix LV is attacking the problem from the other direction: "what if server side rendering was reactive?" vs "what if client reactivity, but with server state as src of truth?"
The model reminds me of sync engines, except Phoenix LV is attacking the problem from the other direction: "what if server side rendering was reactive?" vs "what if client reactivity, but with server state as src of truth?"
October 17, 2025 at 10:44 AM
Someone explain Elixir / Phoenix LiveView to me. What do you like about it?
The model reminds me of sync engines, except Phoenix LV is attacking the problem from the other direction: "what if server side rendering was reactive?" vs "what if client reactivity, but with server state as src of truth?"
The model reminds me of sync engines, except Phoenix LV is attacking the problem from the other direction: "what if server side rendering was reactive?" vs "what if client reactivity, but with server state as src of truth?"
Anybody managed to get a SvelteKit app in a Deno monorepo running on Deno Deploy?
It can't seem to figure out that the root isn't where the dist is, and there doesn't seem to be any obvious way to override the build directory.
Tried many things/am abt ready to throw my hands up and go back to CF.
It can't seem to figure out that the root isn't where the dist is, and there doesn't seem to be any obvious way to override the build directory.
Tried many things/am abt ready to throw my hands up and go back to CF.
October 16, 2025 at 1:01 AM
Anybody managed to get a SvelteKit app in a Deno monorepo running on Deno Deploy?
It can't seem to figure out that the root isn't where the dist is, and there doesn't seem to be any obvious way to override the build directory.
Tried many things/am abt ready to throw my hands up and go back to CF.
It can't seem to figure out that the root isn't where the dist is, and there doesn't seem to be any obvious way to override the build directory.
Tried many things/am abt ready to throw my hands up and go back to CF.