@theo If I used Claude Code only for programming, I would have cancelled it a long time ago. For pure programming Codex is currently much better imo.
English
Piotr Książek
2.3K posts
Piotr Książek
@PiotrKTech
Co-founder Nasesje | Building AI-powered EdTech for STEM students | ex @Autodesk | CS grad | 🇵🇱
Kraków, Poland Katılım Ocak 2025
182 Takip Edilen212 Takipçiler
For every person who replies with a screenshot of their cancelled Claude Code plan, I will donate $10 to open source.
Theo - t3.gg@theo
I can't help but feel personally burned by the Claude Code changes announced today. We put so much work into wrapping the (atrocious) Claude Agent SDK in T3 Code. It was the ONLY path they supported, so we made it work. It was hell. Now our users are getting their rate limits cut by 40x, despite us doing everything right. I listened to the Claude Code team. I had my issues with their direction, but I trusted them and took them at their word. I will never make that mistake again. Until we see significant change, it is safe to assume any statement from an Anthropic employee is a lie on a timer. The rug will be pulled, no matter how many promises are made beforehand.
English
@ItsAlexhere0 It didn't disappear. Maybe it's not worth to use it personally as your main driver (unless you have ton of compute at home, some guys have like @0xSero), but it has god tier price to quality ratio. If you build apps that use AI wanna give your users "unlimited" AI, deepseek is 🐐
English
@bcherny I was about to migrate all my big and small cron jobs from my OpenClaw and raspberry pi to Claude Routines since I am already paying 200$, until I saw this.
It won't be a feature until this limit rises to a reasonable amount.
English
@0xSero Claude app was unusable for months and now I completely switched from the terminal. It's good now. Good enough.
English
@vlelyavin @0xSero Probably some phd-level-intelligence-too-dangerous-to-be-public-model helped design it
English
@ClaudeDevs Nice, my ultrareview was so far launching 3 separate instances with different prompts and cross checking with codex
English
New in Claude Code: /ultrareview (research preview) runs a fleet of bug-hunting agents in the cloud.
Findings land in the CLI or Desktop automatically. Run it before merging critical changes—auth, data migrations, etc.
Pro and Max users get 3 free reviews through 5/5.
English
@kanavtwt Read a book while you wait for the output. Even when I worked at 5 things at a time I was defaultong to scrolling in wait times, but book is much less harm. Even the opposite, a cure.
English
@robin_faraj @AdamBartas Looks cool, but I wouldnt base visual communication on Claude. Things might change fast in not years but months to come.
English
@kylegawley I get 5000$ worth of compute for 200$ in Claude Code. That's why.
English
claude code is a huge risk.
why lock yourself into one model when you can use any model in cursor?
English
@GokulsVision @chribjel True, but also if it's "just work", not handling users with ai, then Claude Max or Codex can handle ~5000$ worth of tokens for 200$ (subsidised). It can be even cheaper now than using open source models. It won't last forever, but for now it's there.
English
@PiotrKTech @chribjel They’ve also probably never heard of Opensource models.
I can’t imagine anyone spending close to $5000+ a month on ai credits and feeling like they could’ve hired someone to do it.
This isn’t an ai cost issue, it’s a skill issue in using ai
English
We hired a junior developer to write the simple code, so we don't have to spend a ton of money on tokens for those basic/primitive tasks
English
Opus 4.7 uses more thinking tokens, so we've increased rate limits for all subscribers to make up for it. Enjoy!
English
@explorersofai @wachig_wiwossa I love open source, but I also love getting shit done fast. So far Claude and Codex are far beyond everyone else.
English
@wachig_wiwossa you could be right, but at least I can run a open source model on my own terms.
English
Opus this, Opus that.
The Claude worshippers are out of control.
Been using GLM for most of my agentic work and it's amazing.
Fraction of the price.
What is wrong with everyone?
English
@SwapHunterz Yeah, I don't even "copy paste" now, just point it to the right context, but I would be far worse engineer if I didn't have chance to code manually for a few years.
English
@PiotrKTech fair point on the thinking part not changing, but ai makes copy pasting from docs way more seamless now, i still see juniors flubbing basics they could've googled five years ago.
English
Thinking part didn't change. When I was starting (5 years ago) there wasn't AI and ppl were complaining about interviews because "on the job I will just copy paste from docs or find it on stack overflow".
But still even with Claude / Codex you run into bugs or architectural decisions that will drive you to the darkest swamps if you don't know how to code.
Yes I write 90% of my code with Claude, but try to revert the proportions when it comes to thinking. It actually saves time.
English
Trying the Claude desktop app. I shit you not it froze on my first prompt.
English
@zuess05 If it can be cloned in 45 minutes, then it is not SaaS, but toy.
English
