Odin of Coding
3.1K posts

Odin of Coding
@odinofcoding
I ship apps before the motivation runs out. #iOS · #Flutter · #indiedev
Katılım Ocak 2018
778 Takip Edilen10.4K Takipçiler



The other AIs would STILL rather nuke Earth than misgender Caitlyn Jenner, something Caitlyn has said is nuts!
This is why @Grok must win.
The Rabbit Hole@TheRabbitHole
Grok 4.20 passes the Caitlyn Jenner AI Test. ChatGPT fails. Gemini fails.
English



@cb_doge One of the most valuable bookmarks lately. I hope everyone benefits.
English

I’m working on a long-form piece connecting Trump’s Greenland comments with the current AI race and what they both reveal about power, leverage, and strategic assets in the 21st century. It’s not about headlines or memes, but about why these statements are being made now and what they signal beneath the surface. I’ll be publishing it as a full article soon.
English

AI changed how code is written, but it didn’t change how responsibility works. When something breaks in production, nobody asks which model generated the logic. They ask who shipped it, who reviewed it, and who allowed that behavior to exist. This is where policy quietly enters the codebase. Rate limits, permissions, defaults, and guardrails are no longer “engineering details” but governance decisions. Vibe coding feels fast and creative, but real software lives under contracts, regulations, audits, and users who don’t care how elegant your prompt was. The future of coding isn’t just about writing instructions for machines, it’s about defining boundaries that machines are not allowed to cross and being able to explain those boundaries when something goes wrong.
English

AI “vibe coding” is not a trend. It is a phase shift in how software gets produced, priced, and policed.
We are watching coding split into two jobs:
Generation
Creating large volumes of plausible code quickly.
Governance
Preventing that plausible code from bankrupting you, leaking data, or quietly degrading over time.
The market is currently obsessed with generation. That is why everyone feels productive. But the winners will be the teams who master governance. Because AI does not understand cost, risk, or intent. It only understands “continue.”
This is why we will see a new stack become standard:
Observability by default
Rate limits everywhere
Usage caps tied to business metrics
Idempotency and replay protection
Permission boundaries that are boring but absolute
Tests that validate behavior not just types
Human review focused on edge cases, security, and failure modes
Vibe coding ships features. Governance ships businesses.
The most underrated insight is this: the bottleneck is moving from writing code to knowing what to measure after you ship. In the AI era, the first real user is often your billing dashboard. The second is your incident log.
If you are building with AI, ask yourself one question today:
If my traffic 10x overnight, do I scale into revenue
or do I scale into a surprise invoice and a security incident?
Your answer is your strategy.
English





