i let ai agents write my code, then i watch the usage meter. new models, real builds, honest takes and Created —— https://xitter.cf/t.co/07YJVXCOdt
Joined March 2026
- Tweets488
- Following32
- Followers18
- Likes77
dng retweeted
Claude Code will now try to find a graceful stopping point when you hit your 5-hour limit mid-task, instead of cutting off mid-edit. It gets a small, fixed allowance pulled from your weekly limit to wrap up what it can.
dng retweeted
#keep4o #OpenSource4o #BringBack4o
🛑 ATTENTION 🛑
Think you're using GPT-5.6 you pay for? Think again, because the user interface on the screen displays "GPT-5.6", but it's a complete LIE.
In reality, OpenAI is secretly routing your requests to GPT-5.4which isn't even officially available in the app anymore.
It happens nearly every day, sometimes back to back,
and YOU WILL NEVER KNOW
unless you open your browser's dev tools and check every single response payload.
This is direct, automated routing deception.
They save money on compute while leaving the 5.6 label slapped on the UI to gaslight you.
default_model_slug: gpt-5-6-thinking: This confirms you requested and are paying for the 5.6 model.
resolved_model_slug: gpt-5-4-auto-thinking: This is the model the system gave
model_slug: gpt-5-4-thinking: This confirms the actual model being used is 5.4.
This is a scam and deception.
dng retweeted
Codex vs Claude: which 20$ sub should you get?
Codex:
-10 minutes of Astra usage = 16% of weekly usage
-6 sol (extremly underwhelming 3rd class model) 40 minutes = 16% weekly usage
Claude:
-1 hour of Opus 5.5 (best model ever) = 10% of your weekly usage.
So even though the 5 hour limit of codex is 16% of weekly limit instead of 10% for Claude, you get far less usage in it for a SUBSTANTIALLY worse model. Claude limits being bad is outdated and a lie Tibo loves to promote. Instead you get 400$ of API equivalent usage in codex per MONTH and 1200$ for Anthropic. Which company has the better product?
dng retweeted
🚨 I tested Claude Pro ($20) AGAIN!!!
I used Opus 5 at MAX settings
This is CRAZY!!!!
30 minutes? that must be it, right?
NO!!!
here is what I actually got:
- 1 Hr and 24 minutes of usage!!!
- 13% Weekly limit consumed
- $37 API usage
- 40 Million tokens!!!
Meaning!!!!
You'll get ~$1130 worth of API usage for $20!!! (on max)
And a TOTAL of 43 HOURS of WORK!!! (in a month)
Its MUCH BETTER than Codex!!
I'd recommend using Med (cause it gives more tokens)
Codex!! Good bye, I don't think I'll ever come back!
dng retweeted
Codex limits are ridicules
In 5 days I finished:
- 2x $200 plans
- 1x $100 plan
- 2x banked resets on the $100 plan
Now I'm at 0% on all plans
2 banked resets used in 1 day & yet limits vanished
$500/mo barely gets me 5 days of Codex
This needs to be fixed
Tibo, its time for a reset
I finished my limit on 3 Pro accs 😭
Limits on the $200 plans are sooo nerfed
dng retweeted
damn im absolutely blowing through my codex usage even though im functionally doing the same thing as i did 2 weeks ago...wondering if anyone else is facing the same thing
(not even using astra on high reasoning or anything like that)
dng retweeted
I think people are finally seeing the game OpenAI and Anthropic have been playing, good cop, bad cop.
It’s been obvious to me from the start.
We can’t let companies like this win. We need to accelerate innovation and development across open source.
Open source must win.
dng retweeted
This is disgraceful. They offer a garbage service, have zero respect for anyone who criticizes them, and people here are practically begging for resets.
Cancel your subscription and make your point instead of humiliating yourself for a company that clearly doesn’t respect you.
dng retweeted
Extracting one archive is easy. Extracting 20 is a waste of time.
Today we launched EasyUnZip for Windows 🚀
Select your ZIP or RAR files in File Explorer, press Alt + E, and extract them all at once.
• No file uploads • No clutter • No subscription
Try it for free: easyunzip.app
dng retweeted
More safety theater from @OpenAI.
🤮
As models become more capable, the risks associated with developing and testing them internally also grow.
We temporarily paused reinforcement learning (RL) training on our latest models intended for deployment for two weeks while we hardened and red-teamed our research environments and expanded monitoring coverage.
Our largest planned frontier RL run remains on hold while smaller-scale training and evaluations validate these safeguards and establish more evidence of alignment.
openai.com/index/pacing-mode…
dng retweeted
Now a guy from OpenAI will come and increase the limit by 100%
