jason

77 posts

jason banner
jason

jason

@frenjason

Sharing https://t.co/4oigowko5j Joining @Livekit Building https://t.co/cWZMnKmmo8

Brooklyn, NY Katılım Eylül 2013
629 Takip Edilen138 Takipçiler
Perplexity
Perplexity@perplexity_ai·
Today we're releasing Personal Computer. Personal Computer integrates with the Perplexity Mac App for secure orchestration across your local files, native apps, and browser. We’re rolling this out to all Perplexity Max subscribers and everyone on the waitlist starting today.
English
766
1.3K
13.8K
13.4M
Milind S
Milind S@milindlabs·
I took Clicky and I made it 5x Faster So I saw that @FarzaTV was using uses Claude's vision to find UI elements on screen, send a screenshot, wait for coordinates back. It works, but it's slow. I replaced that with OmniParser V2 by @Microsoft which is a local Object detection model trained specifically on UI elements. It runs on-device, detects every button, menu, and icon in 400ms, and gives me pixel-precise coordinates. No API call, no latency, no cost. The green highlights you see around the UI elements is the detection overlay and you can see as I am switching the screen it takes no time to detect and highlight which is pretty neat! With Local models improving by the day, Its the right direction for applications like these. Next up: video-synced tutorials where a YouTube tutorial pauses and waits for you to perform each action in the real app. I am not stopping!
English
78
70
1.5K
71.2K
jason
jason@frenjason·
@OrionReedOne @tldraw imagine this as “time machine” on shape selection, click it and scrub through versions
English
0
0
2
51
Anthony Hobday
Anthony Hobday@hobdaydesign·
If a new designer asked you to suggest ONE thing they could do to quickly improve their interface design skills, what would your answer be?
English
52
2
67
32.8K
jason
jason@frenjason·
@EmyLascan Inter for body. Feels like it’s faster to read, maybe denser
English
0
0
2
21
Emy
Emy@EmyLascan·
Recently I replaced Inter with Geist ← Inter | Geist → Which one?
Emy tweet media
English
59
14
458
68.9K
jason
jason@frenjason·
@alpennec imagine you go swipe to dismiss and it scrolls up instead
English
0
0
1
106
Axel Le Pennec
Axel Le Pennec@alpennec·
In #SwiftUI, is there a way to present a sheet whose height is based on the content size, using detents? And if the content doesn't fit, it would be placed in a scroll view (not enough space/dynamic type size). I want to replicate rounded sheets used by iOS. Any idea?
Axel Le Pennec tweet media
English
26
3
259
87.8K
jason
jason@frenjason·
@kylegawley I'm a designer not a dev but I think it's: 1. Tokens - Define primitives, not styles 2. Visible - Style is in-line with content. no need to reference another file) 3. Transferable - New to the team? No need to learn bespoke classes. 4. Works well with AI code completion
English
0
0
1
42
Kyle Gawley
Kyle Gawley@kylegawley·
I've never understood the appeal of Tailwind The power of CSS is reusable classes Style once → apply anywhere Modifying a Tailwind style means doing a find and replace on 100s of instances of an HTML object What am I missing?
English
616
89
1.6K
706.3K
jason
jason@frenjason·
@rahulchhabra07 there’s another situation where you don’t need scale, clean code; or maintenance, and that’s when you make apps for yourself. the more ppl make apps for themselves, the less they need people to do it for them
English
0
0
0
61
Rahul Chhabra
Rahul Chhabra@rahulchhabra07·
Devin is alright for quick demos that don’t need to be maintained or scaled but all serious projects require lot lot more. Writing code is a trivial part of software engineering. The meat of the job is working w all the constraints that comes around building things w technology w people and for people. The optimal software dev via ai assistance is still a chatgpt interface of chat and ai generated code snippets along with explanation and details. Most people who are flooding twitter feeds by calling it software dev killer are actually telling on themselves and should write more code and build more stuff. Software dev isn’t akin to either self driving or making art in the sense that driving doesn’t involve creativity, just path finding and art doesn’t need to be absolutely precise to be meaningful. It’s wayy harder to automate software developers and extrapolating on one isolated dimension is honestly a common fallacy for most of twitter posters.
Rahul Chhabra tweet media
English
122
91
1.3K
303.7K
jason
jason@frenjason·
@alanlaidlaw Did a prompt experiment recently where one llm was a compressor and other was decompressor. Fed in recently written story, got compact gibberish out (<30% original length). The interpreter was good at retrieving conflict, narrative, characters, and even reconstruct the story.
English
0
0
1
18
Alan Laidlaw
Alan Laidlaw@alanlaidlaw·
What if sci-fi got it backwards. LLMs lead to universal translators which leads an explosion of new natural languages. Not universal or planned languages. Not interface two-steps or code words. Natural, but weird, pidgin dialects.
English
1
2
15
1.1K
jason
jason@frenjason·
@soren_iverson Wait don’t/didn’t Facetime group video do exactly that?
English
0
0
9
1.9K
Soren Iverson
Soren Iverson@soren_iverson·
Google Meet dynamically increase the size of your video based on how much you talk during a meeting
Soren Iverson tweet media
English
84
481
17.3K
783.5K
Tyler Angert
Tyler Angert@tylerangert·
what if you could open *any* folder as your desktop? your desktop should just be a folder visualizer, not tied to one arbitrary folder called "desktop"
English
16
14
301
0
Rob Haisfield
Rob Haisfield@RobertHaisfield·
Just learned that Yoshiki has passed away and so many people seem to love his ideas, I’m gonna thread a few here. An app builder on the iPhone where you start by describing the data structures you want in JSON and then nudge it towards a GUI
yoshiki@yoshikischmitz

I've been jamming on this concept for making data-driven designs. Given some JSON, this app will provide you with an interface to describe how you want each entry styled, allowing you to gradually create a more complicated design. Here I create an airbnb-ish app.

English
9
30
274
59.8K
jason
jason@frenjason·
@brijanp Same reason devs hate “it’s not working”, bc you can’t debug that
English
0
0
0
192
tldraw
tldraw@tldraw·
oh no
GIF
English
9
8
156
18.4K
tldraw
tldraw@tldraw·
arrows
GIF
English
20
34
574
105.4K
Yann LeCun
Yann LeCun@ylecun·
Animals and humans get very smart very quickly with vastly smaller amounts of training data. My money is on new architectures that would learn as efficiently as animals and humans. Using more data (synthetic or not) is a temporary stopgap made necessary by the limitations of our current approaches.
English
311
566
5.4K
3M
Jim Fan
Jim Fan@DrJimFan·
It’s pretty obvious that synthetic data will provide the next trillion high-quality training tokens. I bet most serious LLM groups know this. The key question is how to SUSTAIN the quality and avoid plateauing too soon. The Bitter Lesson by @RichardSSutton continues to guide AI development: there’re only 2 paradigms that scale indefinitely with compute: Learning & Search. It’s true in 2019 at the time of writing, true today, and I bet will hold true till the day we solve AGI. incompleteideas.net/IncIdeas/Bitte…
English
138
287
2.5K
1.6M