CodeGlitch
20.2K posts

CodeGlitch
@codeglitch
Code, AI, Glitch.
Arizona | Matrix Katılım Nisan 2018
407 Takip Edilen2.7K Takipçiler
Sabitlenmiş Tweet

.@AnthropicAI is making no money from Opus, and the 1M context window is killing them even more.
Claude works well for a few weeks, enough to get people in, until it gets dumbed down, on purpose or not.
I cancelled Claude Max and went over to Codex.
Way more reliable.
Brad Groux@BradGroux
Something is up with Claude Code usage today. $200 Claude Max, 0%, 52% to 62%, then 68%, 76% and 84% in 5-hour rolling window in the time it took me to write this tweet. WTF, @AnthropicAI? I'm working on one GitHub PR for regression testing. Not folding proteins to cure cancer.
English

One of the most underrated AI workflows right now:
using code to generate product demos.
Not editing timelines.
Not dragging clips around.
Code.
If your demos are mostly UI, motion, timing, and narration structure, tools like @Remotion get very interesting fast.
Then you layer an agent on top to:
- scaffold scenes
- generate visuals
- structure timing
- reuse components
- ship variations quickly
The real unlock is not “AI makes videos.”
It’s that product demo creation starts behaving like software.

English

@codeglitch @Remotion That’s interesting! You been hiding this from me ? 👀🤣
English

There’s a real business hiding in plain sight:
in-person AI workshops for normal professionals.
Not AI bros.
Not prompt engineers.
Managers, operators, consultants, small business owners.
A lot of people do not want another course.
They want someone in the room showing them:
- what to use
- what to avoid
- what actually saves time
- how to use it at work tomorrow
The opportunity is real.
The trap is thinking the hard part is teaching.
It isn’t.
The hard part is trust.

English


@svpino @Yuchenj_UW You just had to tell it not to do it, overkill 😂
English

@Yuchenj_UW I really hate that Claude does this. I had to write my own skill + hook to prevent it from doing this.
English

@melvynx Taking int account the claude code plan is $20, will take a while for mvp too
English


@leerob So, Sshould I stop using claude code and Codex for Cursor or not!
English

Yep, Composer 2 started from an open-source base! We will do full pretraining in the future.
Only ~1/4 of the compute spent on the final model came from the base, the rest is from our training. This is why evals are very different.
And yes, we are following the license through our inference partner terms.
Fynn@fynnso
was messing with the OpenAI base URL in Cursor and caught this accounts/anysphere/models/kimi-k2p5-rl-0317-s515-fast so composer 2 is just Kimi K2.5 with RL at least rename the model ID
English

@RonPaulBot1234 No math in this case, used it a lot in other projects and never seen this before, except from Chinese LLM 👀

English

@codeglitch Math related? I get them when replies need to write more exotic symbols like in math & physics.. my guess the browser doesn't support some of the unicodes or the the app/UI of chatgpt is limited to UTF-8 & then maps or rounds those unicodes to Chinese characters..
English

My startup was hacked!
I launched my own travel eSIM service, eSIMPal
It started making money, the users were happy, and all was good, but today I woke up to a hacked website
Somebody managed to get three 50 (!) GB eSIMs for Kuwait and Saudi Arabia for free, and we started using them heavily
I wired up Claude, and we discovered the issue: the user could pass a parameter from the client to the server and make the eSIM cost 0 dollars
I fixed the issue and blocked this user, and he only managed to use 5 GB worth of data
The internet is full of sharks, boys – triple test all the payment-related code, make sure different LLMs cross-check each other's work
Now I'm writing code with GPT-5.4 and making Opus 4.6 review everything for vulnerabilities
And my hacker bro, if you are reading this, I'll get you your Saudi eSIM, don't worry
Use the promo code IHACKEDESIMPAL for 10% off and chill

English

@LillyLiCT Looks very real, camera movement is good, and environment interactions!
English

@Lowkeighh for me this became more interesting than I was expecting
now I moved on to using it for research and I am creating plans to build on top of this
English
CodeGlitch retweetledi

What if you monitored oracle microstructure, not price?
RegimeIQ analyzes @PythNetwork feed health:
• confidence width
• update latency
• cross-feed disagreement
... to detect market stress before price breaks.
Deterministic signals for Quant trading bots & risk engines.

English










