Brit

@HeyAliux Source?

The power is in the details

Just added an interactive case study to my site on crafting a near-perfect MacBook notch for the web with Claude iamnoman.com/notch

Ok, so we’re officially in the era of BYOT "bring your own tools" where you can create your contextual tooling based on the design challenges you are dealing with. Just put together a custom 3D sticker app this morning with @omma_ai's AI capabilities (tool by @splinetool) - Uses Three.js + 3D Decal geometry - Stickers designed in Spline - Upload custom GLBs + your own SVGs/PNGs Links in the comments!

@hobdaydesign You could blur and lighten the things behind it maybe. Kinda a stretch

It's a bit odd for a button to expand in place in all directions. Where would this button sit, and how would this animation look in context? e.g. I assume it would cover up some important interface elements.

@hobdaydesign Maybe it could be an empty state button, for a table or list

Someone should show these people shaders

Ads in figma smh

@mikethree @skirano Because they’re not using magic path ™️

@skirano please explain how this person is setting the entire industry back by years

Please don't do this. This is setting the design space back years.

@jordwalke Tell us

Few understand what this unlocks. It’s not just about having pretty wrapped text.

@cgtwts I don’t get it

@mil000 Apple hires freelancers and small agencies for lots of stuff. Including ads.

For what exactly? Ads? They already have a team for that. None of this is actual UI

Since we all know that terminals are made for complex UIs... I decided to make T1Code (1T, because a terminal is all you need). I know @theo really likes this kind of complex UI right on the terminal... so lets hope he likes it!

@shiri_shh New startup idea, provide shade from space

This startup lets you ORDER SUNLIGHT from space to your exact location in 30 seconds 😭

someone reminded me that i haven't shared details abt my process for the draw with claude project yet. i put together an overview here! lelezhang.design/drawing

youve seen light mode youve seen dark mode but have you ever seen sunny mode?

@josesaezmerino Screens are wider tha they are tall, lists are easier to scan when they’re vertical, And you can easily hide the left sidebar

I don't understand vertical tabs. Why would I want less space for the website content I actually want to look at?

@iamarunabh Idk about this being good, the springiness would be more impactful when it does show up if this design had the corrected `font-variant-numeric: tabular-nums` effect

Like this x.com/iamarunabh/sta…

If you break the chars and add spring transition, the moving text would look really pretty.

@pizzaboy @DannPetty

@DannPetty I spoke about this a few years back (it was having the ability to context shift as a manager but same thing), people think it's in-and-out but it's the opposite

What non-doers don’t understand about meetings… A 30 min meeting is actually almost 1.5 hrs of time. Mind shift before, the meeting, mind shift after. For any mtg, make sure it’s worth losing 1-2 hrs of getting things done time. Daily meetings are bad for your team.

@MarioBrothBlog I read “crashing in the game” lol

In Mario Kart Wii, crashing the game and activating the crash handler will eventually start playing back the contents of the Wii's RAM as sound. Since the data was never intended to be interpreted like this, it ends up sounding like experimental harsh noise music.

@GBminA @androolloyd @steipete No shit, current infrastructure is worse than the next version of infrastructure, that doesn’t mean the previous versions weren’t stepping stones to the goal. Superiority complex much?

I’m spending my time on this. Yes, LLMs can make mistakes and make things up—no one can argue otherwise at this point—but here’s the fundamental question I’m asking: Why haven’t we been able to create a flawless model? Because, in my view, our starting point was wrong. We’ve always tried to make a model resemble a human. Even a superhuman makes mistakes. So why are we training LLMs by using humans as examples? Why do we want them to write code like humans? At this very point, human—and therefore LLM—errors are, of course, inevitable. For example, AGI will never exist—at least not with the current LLM infrastructure we’ve modeled after humans; it’s impossible. A human cannot be an AGI; the starting point is wrong. So why did this happen? Why are we going through this process? Because control and trust mechanisms are emerging. Because the LLM behaves like a human, we can see and understand what it’s doing. We also see its mistakes—especially when it says, “I used the wrong tool, I’m sorry,” “I misunderstood, I’m sorry,” or, without explanation, “I made a mistake, I’m sorry.” Okay, you apologized, but in the meantime, you spent my million tokens—what are we going to do about that? Also, since this is a proven mistake, replace the tokens you spent along with your apology, and fix my limit as well. So, how does that sound?

This guy emailed me asking for a *token session refund* because his claw made mistakes. 🙃