Rob Pruzan

21.7K posts

Rob Pruzan

Rob Pruzan

@RobKnight__

Building Zenbu | prev Next.js team, react scan co-author

San Francisco Katılım Ağustos 2018
1.1K Takip Edilen4.6K Takipçiler
zeb
zeb@zebassembly·
so what do i name my Vite-based coding agent app? It's built with around a Vite dev server to allow for hot-reloads and powerful extensibility/self-modication I was going to name it aixt but npm thinks it's too close to next 🥲
zeb tweet media
English
11
3
88
9.4K
Rob Pruzan
Rob Pruzan@RobKnight__·
@Qromerolauro > - my favorite ChatGPT Desktop app feature (option+space to ask questions to the floating window) is gone, and now i need to look for an alternative raycast 100% for this
English
1
0
1
98
Rob Pruzan
Rob Pruzan@RobKnight__·
@emollick This also happens with opus if you don’t clean up comments it leaves in the codebase
English
0
0
0
153
Ethan Mollick
Ethan Mollick@emollick·
My one serious piece of advice having used Fable a bunch before release is that, unless you are careful it develops its own internal bizarre cadence & dialogue over long tasks. If you aren't asking it report in plain language, this starts to creep into everything, including menus
Ethan Mollick@emollick

One thing I mentioned only in passing in my Fable post is that, for long running tasks, Fable starts to develop its own dialect as its many agents and tasks reinforce themselves and make Claudish language ever more Claudish. You need to ask it to report out in plain English.

English
56
26
650
80.8K
f1shy-dev
f1shy-dev@vishyfishy2·
i ain’t ever seen a waymo cheaper than a uber
English
1
0
5
549
Rob Pruzan
Rob Pruzan@RobKnight__·
@kohjingyu agents in slack in general (claude tag/cappy/devin etc) is rlly ergonomic, especially on mobile
English
0
0
1
204
Jing Yu Koh
Jing Yu Koh@kohjingyu·
If you work alone, is there any point in using Claude Tag instead of the CC/Codex apps?
Jing Yu Koh tweet media
English
1
0
20
1.6K
Christian
Christian@_heyglassy·
Today is my first day @PlanetScale as a Developer Educator. It hasn’t quite sunk in that I work at the database hat company I’ve been a fan of for years. PlanetScale is the gold standard for infrastructure companies; there have been many days when I’ve worn the database hat and tried to emulate their spirit in my work. I’m excited to join this set of heretical, truly p99 folks. You could spend an hour bumping into people at the office and learn that the keyboard rack was handmade by an engineer who flew in from the East Coast to install it, or that your office manager is a sake sommelier with a really granular rating spreadsheet, and that wouldn’t even scratch the surface of how great every single person is. I always have a great time nerding out about databases, and I’m excited to bring more of that joy to you through events and content, or over coffee. If you want to chat, please say hi in my DMs or at christian at planetscale dot com.
Christian tweet media
English
112
8
637
59.8K
Ian Livingstone
Ian Livingstone@ianlivingstone·
I want to mount Notion to my file system - when ?
English
11
0
13
5.3K
Rob Pruzan
Rob Pruzan@RobKnight__·
Nobody talks about how hard it is to get good emojis into a new slack
English
8
0
15
2.4K
Rob Pruzan
Rob Pruzan@RobKnight__·
@ksw_arman i would prob pay for an agent that can consistently source new relevant emojis
English
1
0
1
75
arman
arman@ksw_arman·
@robpruzan dude its so hard. idk why theres no good app for this yet
English
1
0
2
261
Rhys
Rhys@RhysSullivan·
there's this valley of despair when programming with agents you either have to be an inference maximalist or minimalist, but anywhere in between you just get slop and pain
Rhys tweet media
English
28
15
284
88.4K
Lucas Meijer
Lucas Meijer@lucasmeijer·
I'm now running pi with a web frontend "out of the box", but I still want my project-based extension power. shall I trust the in-sandbox extension code? feels like a pretty big leak. move agent into sandbox again? also sadface, extensions should be able to augment my web ui
English
6
2
17
4.4K
dax
dax@thdxr·
can someone train an LLM only on data from before like 1600 or something would it sound like a person from back then
English
207
28
2.9K
261.5K
Rob Pruzan
Rob Pruzan@RobKnight__·
@RhysSullivan @ksw_arman True but also I prefer the agent to read the actual commit content when searching through git history
English
0
0
0
23
Rhys
Rhys@RhysSullivan·
@robpruzan @ksw_arman i lowkey like their commit messages because of how much context it captures
English
1
0
0
150
arman
arman@ksw_arman·
am i the only one who hates using agents to commit/push changes? it takes like three commands (add, commit, push), and its so much faster for me to just type them myself.
English
10
0
20
2.7K
Rob Pruzan
Rob Pruzan@RobKnight__·
@RhysSullivan Really hype about plugin systems/patchable software in an inference loop future, since it sort of lets you push computation to the client by letting users run loops on the product for you (e.g they want a new feature, agent implements it for them locally)
English
0
0
2
99
Rhys
Rhys@RhysSullivan·
for what it is worth though, inference loops is where all of this is going to end up for the most part and this tweet from peter is the equivalent of what the vibe coding tweet from kaparthy was a year ago
English
1
0
9
1.6K
Rhys
Rhys@RhysSullivan·
twitter's best worst quality is that it forces all takes into an extreme you're either a tokenmaxxing inferencel looping your agents or you're a promptchud the truth lies somewhere in between you absolutely should have agents that can reproduce problems in your products, look for optimizations, automatically open PRs on issues software has never been perfect, even before AI bugs still made it to production, the more of these you can catch and can fix automatically the better at the same time, the underlying engineering still does matter - i just spent 3 days fixing some bad product design and architecture that i originally tried to defer to agents to implement if i wasn't looking at the code, or i was just trusting loops to fix it, my product would be stunted in quality because i would've just kept digging myself into a worse path anyways, setup your loops but prompting your agents is also fine - get the balance
Peter Steinberger 🦞@steipete

Here’s your monthly reminder that you shouldn’t be prompting coding agents anymore. You should be designing loops that prompt your agents.

English
50
33
700
58.9K
BinBin
BinBin@binsquares·
smolvm has hit stable release: v1.0.0! You can now fork smolvm. It means you can fork to create virtual machines off of an existing one in less than 100ms, with all the processes cloned and running. smolvm is the first to have this feature + cross platform compatibility (macOS and linux natively). Here's a demo of a counter continuing on a forked clone while I only started it on the original!
English
31
24
285
55.7K
Rob Pruzan
Rob Pruzan@RobKnight__·
Zenbu is a part of YC!! We are building a hyper extensible IDE for managing coding agents Our goal with Zenbu is to bring neovim/emacs level extensibility into this new class of "IDE" You can try it today at zenbu.dev
Y Combinator@ycombinator

Zenbu (@zenbu_labs) is an extensible IDE for coding agents. Run agents in parallel, manage their work, and customize the IDE with plugins. Congrats on the launch, @robpruzan! ycombinator.com/launches/Qey-z…

English
9
3
32
2.4K