I do SWE-ing for a living. My opinions are my own.
Universe, Milky way, Earth
- Tweets 6,313
- Following 156
- Followers 103
- Likes 39,046
why tf is google cloud so useless? I have 3 projects (had more, deleted them) and I still cant add new projects @googlecloud wtf is this? I shouldnt have to request and wait for more limits. This is so dumb
1
1
138
Everytime I see similar idiotic estimates, my braincells die from cringe 😬
This RESET will cost OpenAI $125 Billion.
They have 50m active paying users for ChatGPT, most users spend $200 monthly, which means $50/week on average.
So one reset price is $50/customer.
They provide 50x more usage for the price you pay, so $2500/customer in inference for one week of usage.
That brings total of $125b total for 30m users.
Kudos to OpenAI. Their investors will probably curse them for this.
24
Kinda crazy to me that developers think prompting AI is engineering :) AI is reading the code, finding bugs, finding the optimal solution, creating a PR and later merging you it. Giving away all your reasoning, architecting abilities to an AI. But sure you're the one doing engineering, right :)
The word "engineering" has become so cheap that people nowadays use it to feel good about themselves.
Don't get me wrong, I'm not anti AI but you're just fancy manager or an orchestrator at an opening. You just have an intern who's really smart and better at engineering than you are and you're just giving them instructions. A meat proxy.
11
Tldr; I have skill issues
I always accidentally press some random key combination that makes the program do something unexpected (and usually something I cannot revert).
Never in my life I wanted to do something with the keyboard apart from typing. I know that people think they are faster using only the keyboard. They are not, I will die on this hill.
But the even bigger problem is this: x.lingyaoai.com/aramh/status/210561427…
1
7
Omg I asked it to replace it with tanstack pacer. This mf is taking 10 mins to do so. DUDE IT'S JUST A BUN INSTALL AND GIT CHECKOUT 😭
I've been maining Claude and opencode for some time. I forgot how slow codex is... Yikes. Asked both 5.6 and 6 Luna to add a denounce utility to the repo. It's been 11 minutes. No real progress :)
What happened to codex man
35
I've been maining Claude and opencode for some time. I forgot how slow codex is... Yikes. Asked both 5.6 and 6 Luna to add a denounce utility to the repo. It's been 11 minutes. No real progress :)
What happened to codex man
1
76
Coding agents are farrrrrr better that chatapps. Chatapps will keep talking to your in a condescending tone and give you generic tips. While simultaneously eating up your limits 🙃
22
The Doctor retweeted
Hey @theo, No need to fork, just turn off the AI features, it is built in:
"chat.disableAIFeatures": true
Anybody have a VS Code fork with all of the extra features and AI stuff removed?
I want a minimal editor that doesn't spam me with requests to use a mediocre agent when I just want to quickly look at some code changes.
10
5
394
80,680
The Doctor retweeted
Dude!
We always knew Rust is faster.
But we chose rails for different purposes.
If you're done with rails, hand over the trademark and let the "losers" like me continue with what we actually love.
And please don't come to next year's conf, esp don't do the keynote.
Enjoy your unlimited tokens. The rest of us still have to deal with reality.
Been running a beta version of Campfire rewritten in Rust today! The code itself is still ugly as sin, and 6x as verbose, but the conversion was free*, and I never had to look at the Rust directly, so this is awesome 🤩 github.com/basecamp/once-cam…
48
43
1,101
168,398
Gpt 6 astra is going through the same phase as gpt 5 did last year lol. Remember how quickly it became irrelevant
2
12
The Doctor retweeted
It's funny how rebuilding basic utilities using modern technologies has become a new kind of dick measuring contest. It's good though. We're gonna get a lot of good tools. But it's really funny.
made a native macOS one (Rust + Swift)
→ 4.7 MB app
→ 278k files in 0.7 s
→ whole Mac (4M files) in ~13 s
→ 1.4× faster than disktree, half the memory
1
1
6
437
Can someone explain why I am getting far less usage while others are getting virtually unlimited usage :) @AnthropicAI @lydiahallie
1
1
42
Looks like I'll be turning one of my hetzners vps into a dev box with t3code. How's everyone using their devbox? Are you guys just logging into your vps, enable a systemd service for t3code, add git, claude credentials? Or are you containerizing it?
30
People who knows nothing about tech should kindly stfu about tech :))
OpenAI and anthropic are casually exploiting companies but sure oss models are bad
We're used to thinking of open-source models as an unadulterated good. But in the case of AI, they can actually pose additional dangers, as @ReidHoffman and I got into at #CGI2026. I appreciated this nuanced discussion.
62