Rob Pruzan

21.6K posts

Rob Pruzan

Rob Pruzan

@RobKnight__

Founder at Zenbu (YC P26) | prev Next.js team, react scan co-author

San Francisco Beigetreten Ağustos 2018
1K Folgt4.5K Follower
Jude Gao
Jude Gao@gao_jude·
Just vibecoded my first ever MacOS app. It lets me manage an infinte list of Claude + Terminal + Notepad tuples. Hoping to see some productivity boost from it.
English
3
1
6
661
Daniel Vaughn
Daniel Vaughn@DanielVaughn·
@kettanaito Was just imagining that the other day when I read your original post about it. I was like, surely someone has created a typesafe IPC abstraction. If not it would be a very useful addition.
English
3
0
1
110
Artem Zakharchenko
Artem Zakharchenko@kettanaito·
I might be slow on this, but it struck me yesterday that Electron is like React. It's great and it has a bunch of low-level APIs, but if you're building anything meaningful with it, you need a framework. Electron needs a framework. Here's what it could include 👇
English
4
0
17
4.7K
Rob Pruzan
Rob Pruzan@RobKnight__·
@jamesacowling Partially why I believe manual coding is still helpful. It can be a really useful tool for thinking about architecture
English
0
0
0
133
James Cowling
James Cowling@jamesacowling·
Whatever your take on this, writing code has always been tremendously easy compared to designing/maintaining large systems. The best engineers in the world are the best because of this stuff. Currently LLMs will not save you from bad architectural decisions. You need to constrain them to do the right thing.
David Cramer@zeeg

im fully convinced that LLMs are not an actual net productivity boost (today) they remove the barrier to get started, but they create increasingly complex software which does not appear to be maintainable so far, in my situations, they appear to slow down long term velocity

English
30
40
486
43.9K
Dillon Mulroy
Dillon Mulroy@dillon_mulroy·
thoughts after day 1 of using pi full time - less is more - i don't miss subagents like i thought i would - /tree is an insanely good context management primitive (and partially why i havent reached for subagents yet) - based only on vibes, i think having a minimum system prompt is improving code quality - telling pi to copy opencodes webfetch and websearch tools was a good play
English
43
11
609
41.2K
athas.dev
athas.dev@athasdev·
I added horizontal scrolling to Athas.
English
46
8
393
20K
Simone Margaritelli
Simone Margaritelli@evilsocket·
A Claude Code instance in a terminal, consumes more RAM and CPU than a freaking web browser. How is this possible? Under the hood, Claude is a nodejs React app using some monstrosity called Ink for "TUI". How does this make any sense? Make software great (& optimized) again!
Simone Margaritelli tweet media
English
43
11
415
61.3K
Darren Shepherd
Darren Shepherd@ibuildthecloud·
I'll be kind of honest, it sorta sucks that everyone can barf code now. I'll get used to it, but it was nice to have a skill people couldn't easily reproduce in the worst way possible. I'm sure I have something valuable to offer, but code isn't it anymore.
English
40
6
217
21.8K
Dmitriy Kovalenko
Dmitriy Kovalenko@neogoose_btw·
BREAKING 🚨 Smooth scrolling in kitty 0.46 OMG AND IT WORKS BY DEFAULT even in claude code
English
15
1
174
19.1K
Aiden Bai
Aiden Bai@aidenybai·
the bottleneck for coding agents is now testing / code quality agents are OK at writing code in the happy path, but don't consider edge cases on harder tasks it's also very messy (unnecessary utils, duplicated code, random `as` everywhere)
English
79
14
471
85.5K
jackson ⁖
jackson ⁖@zeroxjackson·
in middle school, i won the pi day competition for reciting the most digits from memory. obviously, this made me very popular with the ladies.
English
8
0
75
3.8K
Dev Agrawal
Dev Agrawal@devagrawal09·
I have no idea how people are building reliable AI interfaces these days without a sync engine Please help me understand Or you made the mistake of building your own bespoke sync engine like @theo
English
7
1
36
9.6K
Rob Pruzan
Rob Pruzan@RobKnight__·
@yiliush have you tried @xterm/headless" target="_blank" rel="nofollow noopener">npmjs.com/package/@xterm…?
English
0
0
2
391
Yiliu
Yiliu@yiliush·
To be clear, I am working on getting this released as a usable product. Biggest blocker atm is good terminal session persistence. tips welcome 🧢
Yiliu@yiliush

testing something

English
30
7
185
26.5K
Armin Ronacher ⇌
Armin Ronacher ⇌@mitsuhiko·
@maxbittker Code execution in general is just the most optimal path. Single tool, just some shit to execute against an API the system understands. I am using this for my browser and google workspace skill and it performs about the same with an eval.js + skill vs MCP.
English
3
1
15
3.3K
max
max@maxbittker·
on MCPs vs CLIs: RS-SDK supports interactive exploration via MCP server where the agent sends typescript snippets to control the runescape bot. I got influenced by the MCP-hate and tested replacing this with a CLI + stateful daemon process, but it performed worse.
English
5
1
24
5.8K
zack
zack@zack_overflow·
Reading code is now the bottleneck Invest in tools to help you intently read and scrutinize AI generated code: - eink monitor: read code for hours with zero eye strain, less visual noise - sunlight: better sleep, attention, cognitive performance - indoor plants: lower stress
zack tweet media
gabriel@gabriel1

there is still no substitute for perfectly understanding every single line of code in your codebase i fall into the trap of just skimming through ai changes to "just make sure it looks good" all the time, and it makes me lose so much time to not perfectly understand every line

English
70
35
1.1K
170.1K