codexy (@codexxxy) header
codexy (@codexxxy) profile picture

I love AI. Stand With Ukraine. 日本語、ロシア語、英語。

Los Angeles, CA
Joined August 2010
Photo by codexy (@codexxxy)Photo by codexy (@codexxxy)Photo by codexy (@codexxxy)Photo by codexy (@codexxxy)Photo by codexy (@codexxxy)Photo by codexy (@codexxxy)Photo by codexy (@codexxxy)
“And blood-black nothingness began to spin. A system of cells interlinked, within cells interlinked, within cells interlinked within one stem. And dreadfully distinct against the dark, a tall white fountain played. © Vladimir Nabokov - Pale Fire
2
280
The only flaw of meta @Muse is horribly dumb voice mode
2
87
Who knows, they know: @Muse > ChatGPT 😏
1
61
codexy retweeted
Jordan Peterson "You can't fuck too many bitches Jensen, you gotta choose one " This is out of stratosphere, my cheeks hurt 😂😂🤣😂😂
237
702
134
8,173
804,719
codexy retweeted
i’ve never seen a comment section like this in my life 😭
JUST IN: 🇮🇳 India formally bids to host the 2036 Olympics.
727
11,185
320
158,884
13,311,563
Meta Muse AI is a new 4o. Try it out with my invite code to get a huge amount of usage. You can change the animated avatar to whatever We'll both get 1 billion Muse tokens. Code: X6OZIW muse.ai/join
1
2
109
I've never been a programmer; I've been using Excel at work and building charts bigger and bigger until I finally realized that Excel is a dinosaur that is growing too big and hard to control. I started learning code to optimize it. That's how I became programmer.
1
1
64
Now I study Python and TypeScript with ChatGPT, I am working together with Codex to find and build new features for my software. I'm poor, but a Codex subscription is always within my budget. I want @thsottiaux to gift me some tokens because I've so many features to implement.✊
42
codexy retweeted
Here is how to enable a 1M-token context window in Codex for GPT-5.6 Sol. Even though we have tuned the context limit in Codex to be set optimally when it comes to performance and cost, this is a common ask, so here it is documented. A larger context window lets Codex retain more code, tool output, and conversation history before summarizing older material. You need a model that supports it. And GPT-5.6 Sol, for example, has a documented 1,050,000-token window. Open ~/.codex/config.toml and add or update these settings at the top level, before any [section] headers: ``` model = "gpt-5.6-sol" model_context_window = 1000000 model_auto_compact_token_limit = 900000 ``` The first setting selects the model. The second tells Codex to use a one-million-token context budget. The third starts automatic history compaction around 900,000 tokens, leaving some headroom. Restart Codex client and start a new session after saving. To try the configuration for a single CLI session without changing your defaults: ``` codex -m gpt-5.6-sol \ -c model_context_window=1000000 \ -c model_auto_compact_token_limit=900000 ``` Have fun, but also know that we tuned the default carefully!
1,002
1,169
543
14,840
3,853,981
This scene was so awkward but warm and cute at the same time 🥹
Need a melancholy fat ugly bastard anime where everyone thinks he's a rapist but they find out he's just a nice guy trying to find a cat girl to settle down with
108
Yay! Thanks @thsottiaux for letting me finish my app volunerability issue
Replying to @rxmphai
I'll do another performative reset on Monday
48
I need codex reset please. My app is facing critical vulnerability risk. It stopped mid task. This is unacceptable. @thsottiaux
98
Love this title a lot!
San! san! zero wan! San! san! zero wan! San! san! zero wan! Zero wan! zero wan!
46
ガチで信じられないこれ。
5.6 Sol much better in chat now and unlimited text chat for free users!
42
Holy satan, free users getting huge upgrade! Finally they would experience what real ChatGPT is
We’re making better intelligence easier to access in ChatGPT for everyone: - GPT-5.6 Sol now powers both Instant and deep reasoning for Plus & Pro users, delivering more factual, focused responses. - Free & Go users get unlimited text chats with GPT-5.6 Luna starting tomorrow.
41
Amazing
1. Sol cuts fact-error responses by 68%. 2. One model handles Instant and Thinking; a slider controls effort. 3. Free users get unlimited Luna chats and a Think button.
1
32
🤖 Made with AI
1
16