scumfuck flowergirl // i know a little bit about computers // priv: @Zbuildstd
17
- Tweets 20,465
- Following 1,793
- Followers 5,783
- Likes 53,834
opencode go’s glm-5.3-flash has abhorrent caching. half my requests cache properly and the other half have literally 0 tokens cached. i am sending the opencode header properly. what gives??? it’s costing a lot of usage
9
156
7,741
the prophecy foretold your existence, le chaton fat
Meet Mistral Large 4, aka Le Chonk.
• 1T parameters, natively multimodal. 49B active.
It is the best open weights model from US or Europe on aggregated benchmarks.
• State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding.
• Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure.
• Available to all via API today. Working with cybersecurity partners privately.
Open weights release end of October.
6
5
289
7,451
nullptr 🐱🍩 retweeted
wor$t man in america. out 12/4.
presave and preorder now slayyyter.lnk.to/wmia
989
11,476
48,731
1,284,379
nullptr 🐱🍩 retweeted
Capy is already the coding agent with the best UX by far, and using it as a desktop app is soooo nice. Super happy with it, been using it for a few weeks.
Introducing Capy Desktop: the best way to orchestrate coding agents across all your devices
Remote control and access local files/shells from any machine
Close your laptop and Capy continues in the cloud
Now available for MacOS, Windows, and Linux
Comment for $100 in credits
2
2
15
695
nullptr 🐱🍩 retweeted
Introducing Capy Desktop: the best way to orchestrate coding agents across all your devices
Remote control and access local files/shells from any machine
Close your laptop and Capy continues in the cloud
Now available for MacOS, Windows, and Linux
Comment for $100 in credits
526
95
780
98,380
lol i did all of this yesterday and told a few of my friends about it since i knew someone would tweet about it and ruin it. shame
GUYS they are giving away FREE plushies and stickers!
Go to claude.dev/terminal
type /plushies or /stickers
go get yours fast, they might run out pretty soon!
12
2
292
16,785
i really hope qwen4.0 27b has CSA2 or something similar so KV cache becomes essentially free. deepseek can fit 1M context in literally 900MB, imagine how useful that'd be for models which already push the vram ceiling for most gpus
7
2
134
2,817
Before you go all out with your AI psychosis and get the $500 a month plan. You do not need it.
You are doing absolutely nothing of value that requires that amount of money, and if you do, a company you work for is paying for it.
You shouldn't, you don't make any money using the AI's, you produce slop on daily basis and post benchmarks on Twitter for 4 likes and expect a return.
After posting the slop you end up in the perpetually slop making hoping for the next banger of the 8 likes post, eventually you will get consumed by the slop thinking it is the future, all your thinking has been moved over to AI, you need to figure out what to do? You ask AI.
You need to watch your neighbors dog, you ask the AI what to do. Your kid needs help with the homework, you get an agent to answer their questions while you wait for your next slopbench to finish. In the meantime you watch your favorite AI generated streamer.
You have been consumed by slop, you have spent 20 thousand dollars on AI in 3 months instead of just getting a computer that could be running all of this locally at fraction of the cost, but you bought into the hype, you got FOMO, all your friends have the biggest and largest plans, they have their own company that makes a whopping $0 and they are 7 thousand dollars in debt, but their AI agent has promised them 80x growth.
Wake up.
38
25
432
20,460
nullptr 🐱🍩 retweeted
youre rotating apples in your head, im rotating all my ssh keys. stay actively secure, twin
3
22
416
4,584
