Apoorv Sharma

Never stop looking up 🌌

[Apoorv Sharma @ars2107_]

LFG!

[Indiexa @indiexa]

Here's our walkthrough video :)

[Apoorv Sharma @ars2107_]

@FlyaKiet wait... he ranks 111 on the leaderboard without cloud agents?

[Apoorv Sharma @ars2107_]

omg this is so stupid. what's your quest here? why some humble feedback about a company (which i'm guessing) you don't even work for, making you so upset?

i use railway for my work and for hobby projects, lmk if you do work at railway and you think if you have a successful product, you should stop fixing things

[Apoorv Sharma @ars2107_]

superset is competitive af

[Kent C. Dodds 🐨 @kentcdodds]

If you're not selling a software factory then your software factory is a cost center. Invest, yes, but don't get lost in it

[Apoorv Sharma @ars2107_]

@JustJake @princecaarlo @Railway imo, bad ux. if this happens even if i sign up with the same email i have on github

[Apoorv Sharma @ars2107_]

it's time...

[Thariq @trq212]

we’re thinking of killing plan mode and using the shift+tab hotkey to adjust effort levels

I don’t think the models need plan mode anymore, but if you’re a plan mode diehard would love to get your feedback on why

[Apoorv Sharma @ars2107_]

@FlyaKiet @superset_sh @AppStore whoa, congrats!

[Apoorv Sharma @ars2107_]

if this is what pacing looks like, i'm so hyped for the future

[Apoorv Sharma @ars2107_]

@theo 5 claude, and only 3 codex? @sama was right, huh?

[Apoorv Sharma @ars2107_]

@mitsuhiko some people are replying with claude sdk as a solution, lmao 😭

[Apoorv Sharma @ars2107_]

grok 4.7 can build a jeep for you if you ask nicely

go catch grok bots

[Apoorv Sharma @ars2107_]

lmao, 1500+ words banger

[dax @thdxr]

kit finally finished the blog post

he covers how OpenCode 2.0 is able to modify everything about itself in real time even during running sessions

btw post whatever easter eggs you're able to find

https://t.co/7WZOaygFjK

[Apoorv Sharma @ars2107_]

@thdxr buy low sell high

[Apoorv Sharma @ars2107_]

are we finally gonna be blessed with opus 5.5 soon?

[Apoorv Sharma @ars2107_]

@theo skill issue

[Apoorv Sharma @ars2107_]

beautiful

[Apoorv Sharma @ars2107_]

@covacut cool af

[Apoorv Sharma @ars2107_]

@theo x fell off

[Apoorv Sharma @ars2107_]

@RhysSullivan nice

[Apoorv Sharma @ars2107_]

"plan with a dumb model, implement with an intelligent model" is one of the worst takes of all time

in the end, your goal is to get the best possible code from AI, planning with one model and implementing with another doesn't work at all. you'll end up exhausting your tokens on the intelligent model and still end up with total slop

instead, you should divide your tasks based on intelligence. the core parts of your software goes to astra/fable, less important parts of your software goes to models like glm, sol, opus etc

if fable and astra are already expensive for your codebase you could use them for reviews of critical PRs