AI has automated software engineering. What you would expect is that there would be no more work left to do for software. But instead what has happened is that the leverage of doing software has increased so much, that doing anything else is a waste of time
Nobody tells you this: Dopamine from information gathering is a dangerous drug. It’s the dopamine from reading, planning, or learning, but never doing. Stop looking for more information and start acting on the information you already have. Get your dopamine from action.
“the most surprising thing is the lack of public recognition of how close we are to the end of the exponential.”
dario’s predictions have been uncannily accurate so far, and i agree.
it’s 'absolutely wild that people are still focused on hot-button political issues' while we’re about to spawn a
“country of geniuses in a data
center” -- (lord dario).
at what point do people finally see this coming?
In the cyberpunk year of 202x, we make illegal art on our massive parallel processing machines. Corpo bounty hunters named "sloprunners" are tasked with taking it down. The arms race is only just starting.
The closer people are to the cutting edge of AI the more they’re having an existential crisis and losing their minds about what the future will hold.
That’s not a coincidence.
I'm 48 and have worked as a software engineer for nearly 30 years. I've grown numb to the Silicon Valley hype machine. My default posture is "meh, we'll see." What I've seen and experienced firsthand in the past two months is not hype. Ignore it at your peril.
Coders right now after OpenAI's Codex safety update:
>Try to debug exploit code
>Get rerouted to baby mode
>Lose 30 minutes explaining to the model why it's allowed
>Give up and use Claude Code
Sound familiar, 4o users? Welcome to the club.
you're writing a CLAUDE dot md? let me guess. "this project uses React with TypeScript."
brother claude can see the tsconfig.json. you wrote 200 lines describing your file tree to an agent that can do `ls`. you explained that ~/projects/to-do-app is a todo app.
the only lines that matter are the ones where your project is weird and you know it.
"run yarn test:unit not npm test."
"don't touch anything in src/legacy/ or three enterprise clients lose their minds."
"the auth middleware is load-bearing, yes all of it, don't be a hero."
that's it, that's the whole file. if claude would've figured it out from reading your code, you're wasting context window. start with nothing. wait for claude to do something wrong. tell claude to remember to not do that. that's your CLAUDE md.
Feels surreal to me that you can still walk around an office in 2026 and see rows of white-collar workers editing spreadsheets, reading docs, writing memos.
This will all be gone in two years, just like the human "calculators" of the 1940s. Do they know it?
vibe coding:
> super pumped about an idea
> “let’s build it babyyyyy”
> bugs keep rolling in
> “was this even a good idea?”
> existential crisis
> panic mode about your future
> “will never escape the underclass”
> bury the project down the grave
Devs who are feeling overwhelmed, take an hour out of your workday and do this:
Setup
1. Get Anthropic Pro ($20), with a plan to upgrade to 5X Max later
2. Download Claude Code
3. Select Opus 4.5 (it's the default)
Loop
1. Start plan mode
2. Plan a small feature
3. Once you're happy with the plan, auto-accept edits
4. Pause the LLM if you're not happy with the output
5. Clear context and repeat for the next feature
Continue doing this until you get a feel for what the LLM can and can't do. It'll take 10-20 hours of practice.