cheaty

8K posts

cheaty

@cheatyyyy

23, sde @ stealth startup, performative nerd

hyderabad Tham gia Şubat 2022

668 Đang theo dõi1.2K Người theo dõi

cheaty@cheatyyyy·6h

this might be the most unnecessary bullshit i've ever cooked up but what can you do say hi to live codex usage in the statusline, ty for cooking with pi-extensions @aliouftw

English

103

cheaty@cheatyyyy·6h

@thdxr how it feels using claude through github copilot for cheaper than a claude sub with all this drama for 3 months lmao (except the 128k context limit kinda blows)

English

10.6K

dax@thdxr·6h

opencode 1.3.0 will no longer autoload the claude max plugin we did our best to convince anthropic to support developer choice but they sent lawyers it's your right to access services however you wish but it is also their right to block whoever they want we can't maintain an official plugin so it's been removed from github and marked deprecated on npm appreciate our partners at openai, github and gitlab who are going the other direction and supporting developer freedom

English

174

288

5.6K

432.5K

cheaty@cheatyyyy·6h

the opencode hype has died right? besides opencode go ofc, that is legitimately a wonderful deal for cheap OSS models hosted well codex harness is amazing for gpt-5.4 especially with subagents being GA i prefer pi for claude or if u want to swap between claude/gpt

English

cheaty@cheatyyyy·7h

@0xSero anything but zed is cooked

English

257

0xSero@0xSero·8h

I had to open this abomination today, holy fuck. How can anyone use this?????

English

162

730

141.9K

cheaty@cheatyyyy·7h

@MandoMichael my goat

English

Michael Mando@MandoMichael·8h

Here's to all u beautiful people who have been down since Vaas. Thank u 💚🦂

DiscussingFilm@DiscussingFilm

First look at Michael Mando as Scorpion in ‘SPIDER-MAN: BRAND NEW DAY’.

English

1.7K

12.3K

203.1K

3.5M

cheaty@cheatyyyy·8h

the x axis is not the problem, the problem is they have only TWO public benchmarks because it's a very general model trained on cursor usage data, so it's good at average/general tasks but will struggle with spiky/jagged workloads (bad at benches), or real cases where it has to fix an unobvious bug in a real codebase

English

5.3K

public_intellectual@publicinte·9h

declining x axis to trick our mind into seeing “top right good”

Cursor@cursor_ai

Composer 2 is now available in Cursor.

English

488

82K

cheaty@cheatyyyy·8h

@badlogicgames they just want their palms to be warm

English

257

Mario Zechner@badlogicgames·9h

just found out my SO is cheating on me

English

7.7K

cheaty đã retweet

Ben (no treats)@andersonbcdefg·12h

OpenAI Newsroom@OpenAINewsroom

We've reached an agreement to acquire Astral. After we close, OpenAI plans for @astral_sh to join our Codex team, with a continued focus on building great tools and advancing the shared mission of making developers more productive. openai.com/index/openai-t…

ZXX

1.8K

83.2K

cheaty@cheatyyyy·10h

@scheminglunatic maybe he likes it when he's behind don't judge

English

alcuin ❄️@scheminglunatic·10h

He shouldn't bother, he is years behind my research.

ThePrimeagen@ThePrimeagen

why is he hard while doing this?

English

1.5K

cheaty@cheatyyyy·10h

@spinelessaisha vitadmin

Español

aisha@spinelessaisha·13h

okay so

aisha@spinelessaisha

i really need to figure out what's making is hard for me to do things, even if i'm doing something that i should enjoy, i just can't seem to focus, and i can only imagine this is bc i'm not fulfilling a need of mine so my brain refuses to cooperate... but what is it

English

319

cheaty@cheatyyyy·10h

@VictorTaelin qwen3.5 9b is the best u have, 20 tps on my macbook air m4

English

565

Taelin@VictorTaelin·11h

So what's currently the best OSS coding model that can run on the base Mac Mini, to be used with Pi? Is there anything decent? Is it Qwen3.5-9B?

English

120

17.8K

cheaty@cheatyyyy·10h

@Presidentlin atleast we have cursorbench x.com/i/status/20346…

Cursor@cursor_ai

Composer 2 is now available in Cursor.

English

Lincoln 🇿🇦@Presidentlin·10h

@cheatyyyy Yeah, it's a stretch.

English

131

Lincoln 🇿🇦@Presidentlin·11h

It does "beat" Opus 4.6 but benchmarks mean so little. It's going to be a good worker model.

Jimmy Apples 🍎/acc@apples_jimmy

Apparently Cursor is going to release a coding model better than opus 4.6 and cheaper as well ( maybe tomorrow ) Can they regularly do this to keep up though ?

English

5.6K

cheaty@cheatyyyy·13h

oh my god it's python versus javascript

OpenAI Newsroom@OpenAINewsroom

English

cheaty@cheatyyyy·15h

@Notesy98 @DanielRPK i doubt it, would just make an already cramped movie even more cramped

English

1.2K

Dylan@Notesy98·16h

@DanielRPK Thoughts on the Man Spider stuff being a red herring and its actually a symbiote, through the hive mind, altering his DNA and powers?

English

8.2K

Daniel Richtman@DanielRPK·18h

Oh? 👀

Inside Fandom@InsideFandom

The difference between Peter’s Web Shooters and his Organic Webbing in ‘Spider-Man: Brand New Day’🕷️

QST

121

5.6K

185.5K

cheaty@cheatyyyy·19h

the problem is that it's not that i dont know how to do exactly what im asking the llm to do using bash commands or literally manually making the folder and dragging things around in the gui, just thinking of what i want my computer to do and sending it is just way simpler and lighter cognitively when i have the confidence it'll do it with no real downside, and now it has the context of the files i might be working with for the next session

English

Lincoln 🇿🇦@Presidentlin·19h

@cheatyyyy Yea I agree hard on this one.

English

Lincoln 🇿🇦@Presidentlin·20h

I've become very lazy as a dev. I ask the model to move my files and do little fixes that would take me like 1 sec to do.

English

454

cheaty@cheatyyyy·1d

@ferologics and this works for the openai-codex provider right?

English

fero@ferologics·1d

honestly, after using it on `low` all day it feels a lot better than default `medium` that Pi hardcodes under the hood. less walls of scrollback to plow through. doesn't completely fix 5.4 for coding for me still, but feel like this helps get closer to what 5.3-codex was.

fero@ferologics

Just published a tiny Pi extension that adds per-model OpenAI text.verbosity control with a keyboard toggle. Set low / medium / high per model, cycle it instantly with Alt+V, and see the current verbosity inline in Pi’s footer. Great if you want the same model to be terse in some workflows and more expansive in others.

English

202

cheaty@cheatyyyy·1d

@kr0der funny

English

301