I use Microsoft Edge and need >134GBs of RAM for my terminal (yeah I vibe code)

Hi! Welcome to Cody’s website.
I use Microsoft Edge and need >134GBs of RAM for my terminal (yeah I vibe code)
OpenAI’s codex
has demonstrated something I’ve struggled to for a while: we’ve been doing AI software engineering wrong. Or at least, there’s a better way with CLIs and files as the interface. What’s old is new again.
In this post, I demonstrate the core workflow for modern agentic software engineering that can be applied to pair-programming or automation via CI/CD:
task.md
file with all relevant out-of-repo contextai "open the task.md and work on it"
gh issue view NUMBER > task.md
and such…Makefile
or similar (I like justfile
) and instructions your ai
on using themYou can then leverage standard software development lifecycle (SDLC) best practices and GitOps for modern agentic software engineering.
I suspect banning advertisement – at least digital advertisement – would make the world a better place.
“FUCK!”
Last week there was a disaster and that disaster has a name.
On Friday I woke up pleasantly surprised to find my power was on. Over 10% of my fellow Floridians – more than a million people – were not so fortunate. At least one was dead with the total toll continually rising.
I took Murphy and Nix on our usual walk to the dog park. Unusually nobody else was there. Had they lost power? Had fallen trees blocked their driveways? Did they have to deal with damage to their cars? Their homes? I didn’t know and at least in some sense did not care. I was content my house was undamaged. A large tree branch almost landed on my car’s windshield but missed by a foot. It was like any other cloudy morning with a slight drizzle. Disaster and death happen all over the world every day and life goes on.
I would like to abolish the term RAG and instead just agree that we should always try to provide models with the appropriate context to provide high quality answers.
This is important. Obfuscating how to use the language boxes with dumb jargon isn’t helpful.
- Cody
Looking at LLMs as chatbots is the same as looking at early computers as calculators. We’re seeing an emergence of a whole new computing paradigm, and it is very early.
- Cody
I’ve first noticed this problem in the startup world where you see famous people that sell books, talk at conferences, and tweet advice to founders, but when you take a closer look, they’ve never done much founding themselves. They’re like the “entrepreneurship” professor that never built a business. They’re experts in the metagame – they’re polished speakers, engaging writers, and thought-leadering tweeters. The problem, though, is that they’re not judged by customers, the market, or nature, instead they’re judged by their peers. I call them metapreneurs.
The bane of my existence:
InvalidRequestError: This model’s maximum context length is 32768 tokens. However, you requested 33013 tokens (31413 in the messages, 1600 in the completion). Please reduce the length of the messages or completion.
Somebody please give me GPT5-64k access.
See part 1.
The bane of my existence:
InvalidRequestError: This model’s maximum context length is 16385 tokens. However, your messages resulted in 28737 tokens. Please reduce the length of the messages.
Somebody please give me GPT4-32k access.
Updates:
See part 2.