RandomYang

Building @lovart_ai. Ex-ByteDance. Visual for fun.

[RandomYang @_RandomYang]

Couldn’t agree more. This slaps hard at everyone who needs a reason for every single thing they do.

[Mitchell Hashimoto @mitchellh]

Mind boggling to me that I can make a thing faster and there's always people that ask "but why?" What kind of mentality is that? The pursuit of excellence does not need justification. Also, I find in so many cases, we can't know the impact of an improvement until we do it.

For example, one I've talked about before: Ghostty's high IO throughput has enabled terminal program (emulator and TUI) fuzzing at a speed thats incomparably fast to prior solutions. This has resulted in upstream patches to resolve issues in popular projects like btop, tmux, and more.

Speed enabled that anecdotally example that lifted the tides of adjacent communities that don't rely on Ghostty technology at all. I didn't predict this.

Make things better because they can be better and let the results naturally play out.

[RandomYang @_RandomYang]

@DavidSHolz AI kills accidental complexity, not essential complexity. the judgment calls don't go away — they just get more concentrated per hour since execution sped up. that's the fatigue. fix: fewer, longer blocks protected for judgment work only.

[RandomYang @_RandomYang]

@rsms @theo Yes, at present, web development can be relatively easily run in a sandbox environment.

[RandomYang @_RandomYang]

@bcherny The skill combination and decision-making priorities of different stages (Pre-PMF, Growth, Mature) and different fields (Consumer vs Enterprise vs Infra) are very different.

[RandomYang @_RandomYang]

@rsms @theo I think this refers to an Agent in a cloud sandbox. Its runtime mirrors your local environment, so you can drive an agent independent of your local setup anytime — including but not limited to software development.

[RandomYang @_RandomYang]

10 design and implementation prompts is:

[RandomYang @_RandomYang]

Better slop.

[Amber Shen @whosamberella]

AI slop in 2025 vs 2026

AI slop in 2025
> blue-purple gradient
> giant hero headline
> em-dash everything
>”I build this in 15 minutes….”
> black landing page with neon accents

AI slop in 2026
> neobrutalist frontpages
> aggressively clashing color palettes
> beige + orange + brown (sorry Claude)
> intentionally ugly buttons
> mixed fonts in one single page

[RandomYang @_RandomYang]

Modify is back.

[RandomYang @_RandomYang]

https://t.co/7Ej1o9xK0W 卒

[RandomYang @_RandomYang]

@varic @punkesh @jeazous Come on, server and storage also count as costs. Don't take things you can easily access for granted.

[RandomYang @_RandomYang]

@punkesh @jeazous Adobe will not offer you the chance to create drafts for free.

[RandomYang @_RandomYang]

@toddsaunders How do you stop it if I look at your design and recreate it purely from memory?

[RandomYang @_RandomYang]

skill : agent
::
function : object

[RandomYang @_RandomYang]

@dreamwieber What you have done and your enthusiasm are fully worthy of this result.

[RandomYang @_RandomYang]

@tszzl The correct answer may be to avoid building the UI.

[Addy Osmani @addyosmani]

https://t.co/0DyFIXJueI

[RandomYang @_RandomYang]

@benswerd memory leak.

[RandomYang @_RandomYang]

The best practices for coding can be fully reused when writing skills.

[Matt Pocock @mattpocockuk]

Long skills are such a red flag to me

- Hard to audit (and therefore, trust)
- Hard to edit (more text, harder to maintain)
- Expensive to run (more text, more tokens)

The shorter the skill, the better IMO

[RandomYang @_RandomYang]

@SHL0MS Monet created the Monet style.

[RandomYang @_RandomYang]

@tobi It's quite interesting to let the skills within an organization be shared and reused to generate compound returns. But how can we tell whether a conversation or a skill is positive valuable content rather than noise?