Opens profile photo
Follow
Click to Follow seflless
Francois Laberge
@seflless
I build creative tools. 🇨🇦/🇺🇸. Side hustle:
BostonJoined December 2007

Francois Laberge’s Tweets

Leptons looks like it’s single allocation at construction instead of per render like Solid but done in Rust. Must be fast as heck, assuming webassembly to JS marshaling is efficient. I’d guess it has the granularity like Solid to patch only exactly which DOM props have changed.
Quote Tweet
The framework Leptos for Rust has created a template on CodeSandbox! If you want to play with a Rust web framework that feels familiar with React, this is a good place to start 😄 codesandbox.io/p/sandbox/lept
Leptos framework running in the CodeSandbox editor
2
Flutter’s getting webassembly support, probably making it the most performant modern cross platform 2D app UX runtime. Yep, that was wordy.
Quote Tweet
💙 Flutter on the web. Tune into #FlutterForward to learn about three key areas of investment on the web: 💻 Wasm support 🔗 Seamless integration with mobile 💥 Powerful new graphics capabilities Watch live 🔴goo.gle/FlutterForward
GIF showing a chart and text that says "Time to load" set against a purple background.
GIF
Val town is the pragmatic instantiation of my favorite ideas in Unison Web, Darklang, & other future of coding languages. I’m very bullish on environments with certain invariants/constraints but using an existing language instead of a new one. JS/TS is the logical choice.
Quote Tweet
🎉 Show HN: Val Town!!!!
Show this thread
Image
2
25
Crying was a total no go. Couldn't get a single tear to show up. These are 2 sets of results. One set has only subtle sadness and realistic faces, the more anguished versions are pretty borked.
Image
Image
Image
Image
2
Show this thread
I started off my study by trying to generate one of those reference grids of human faces. The type you show kids when explaining emotions, like in the first image. But the results were total monstrosities, so I switched to trying to get single faces showing a range of feelings.
Image
Image
Image
1
Show this thread
It's easier to get AI to generate a blood soaked face than generate faces showing strong emotions like crying or anger!? 🧵 Is anyone doing good work with generating emotional expressions with AI? I'm using . I think it's DALLE-2/stable-diffusion, not their limits
Image
Image
1
Show this thread
Full disclosure, I've never put in the proper time to learn Rust. I just fantasy read the documentation and every few years setup the tools and then don't stick to it. Zig felt more comfortable from my experiments with it, perhaps being less divergent from my comfort with C/C++.
1
Show this thread
I know everyone is using Rust for their Webassembly these days, but Zig is looking real good to me for ergonomics & simplicity. I'm ok with my code only running within wasm, so mem safety matters less, no? *expecting to eviscerate me in comments for bringing this up*
3
4
Show this thread
I've been working with Webassembly more lately. I forgot how much much I enjoy working in a systems programming language. After 15+ years doing mostly web dev, I still can't shake the discomfort of how inefficient everything is.
1
16
You know you are too obsessed with infinite-canvas-based creative tools, when you updated your OS for a new app that required it, then updated 2 more devices’ OSs so you could see the UX decisions for each form-factor. apple.com/newsroom/2022/ Interesting so far & odd/quirky too.
6
Rumors are Apple is at least considering letting other browser runtimes on iOS. This would be amazing. It blows my mind how many users in comments actually don't want this to happen. I'm not sure how many people would actually switch browsers given more >10years using Safari.
Quote Tweet
Apple Considering Dropping Requirement for iPhone and iPad Web Browsers to Use Safari’s WebKit Engine macrumors.com/2022/12/14/app by @rsgnl
Image
1
5
I'm working on getting full Typescript type-checking/auto-completing/etc running in @convey_it, integrated with the existing Codemirror editor I found this jewel from Observable's blog. Thought I'd share it with all you web-based programming builders.
2
29
Show this thread
Working a lot with AR/VR/AI, you unintentionally build traditional VFX skills. What is modern movie VFX but mixed reality techniques with better tools and equipment. I should tap into this and make more videos. This insane video brought that to mind:
3
33
This is the most exciting web standards work to me, by far. Highly recommend watching it to learn about potential webassembly evolution. For starters: Module system, large improvements in the marshaling performance hit, and interesting set up for a many tiny processes compose
Quote Tweet
If you struggled to understand the high-level goal behind WebAssembly’s Component Model proposal (like me...), this talk by @luke_wagner is excellent and got me really excited. (N.B.: Component Model has effectively superseded Module Linking) youtube.com/watch?v=phodPL
2
6
…also you need to consider: commenting systems, native vs web, device form-factors, offline (including full asset sync to device), and really all the anthropology of collaboration you should be deeply thinking about.
4
Show this thread
You have way more considerations than what cool CRDT/OT/etc, such as: permissions and modalities (private, org, sub-group, external, internal, public, maybe transient like a meeting ), branching type asynchronous merge (manual or auto). UX to reduce cognitive load is paramount
1
3
Show this thread
Spline is really hitting that Figma for 3D goal post. Their releases lately have been so gooood.
Quote Tweet
Today we are introducing Conditional Logic! 🎲 - Trigger actions based on conditions: distance, values, or state changes. - New collision events to trigger actions when objects hit other objects. - Game Controls actions for idle, move and jump. 🔥Demo: my.spline.design/minihouselogic
Embedded video
0:29
11.9K views
12
There is a lot of really nice global illumination implementations happening right now in WebGL. ’s path tracer ’s screen space global illumination coding up parts of the ReSTIR paper below.
Quote Tweet
I gave it a try to ReSTIR paper by Nvidia, but only the Reservoir bit, not the pdf resampling. It's probably a broken implementation, but it is doing something for sure (it's most effective in 4k monitors and 120 fps): shadertoy.com/view/stVfWc
Image
7
Here's an old prototype I did while waiting for this announcement to support Passthrough AR in browser.
Quote Tweet
Turn your monitor into a canvas & paint on it using Headless VR controllers. Ignore the delay between controller movement & brush strokes displaying. It's the app not 3d tracking I just posted a longer form article on the @convey_it Headless VR concept conveyhq.com/blog/headless-.
Show this thread
Embedded video
0:03
281 views
2
Show this thread
To be clear. ShapeXR is using an official feature for the controllers that now support a pressure sensitive stylus tip that you add to the stem of the controller. It’s not the most elegant solution, but it will help bootstrap AR/VR spatial pen use cases mixed with other controls
1
1
Show this thread
This was one of the more interesting announcement yesterday. ShapeXR here showing the ability of Meta Quest Pro controllers to draw on any surface in AR/VR. Early times & it already looks like it’s being put to good use. this is more where infinite canvases come in
Quote Tweet
The new stylus controllers add an additional dimension to the #creative process. Check out how we have implemented them to facilitate sketching and commenting during the early #design of #AR, #MR, and #VR apps.
Embedded video
0:50
24.6K views
1
18
Show this thread