sounds less pretentious in context.

Tech folks: Why are normies being anti tech now??

Let's have a look:

Degraded trust in news outlets, misinfo networks, things becoming slop (AI or not), shittier conditions due to capitalism on steroids.

Folks have a hard time knowing where to direct their anger. So what is visible is targeted /1

From an aphoristic blog post I'm working on.

Smaller and smaller agents, dedicated to more specific tasks. Taylorism-Fordism for the LLM. Is it unethical? Not nearly as much as the original: an automated facsimile of human gab can be tailored (Taylored) without guilt. Truly we _are_ the petite bourgeois of the engineers.

@mariozechner.at on best practices. trying to keep this in mind as i spend the next 6 months in the slop machine

mariozechner.at/posts/2026-0...

learning tmux... why do we even have terminal emulators lol

most of that time is spent using the ai to do research on repositories and take notes -- gathering context! -- in order to execute on. the execution itself often needs oversight, but i don't miss the research time. reading through foreign codebases was never fun.

obviously once this contract end this pattern ends as well, though i'm hoping to get a nice gpu to run local models on. even with a 5090 tho i'm only running qwen4-instruct or something. plus, even now using enterprise ai, there's a lot of necessary oversight and waiting.

i do have to say that as a parent and somebody who is addicted to making things, having infinite tokens from my job is really, really nice. just run some models overnight or while i'm playing with my daughter. weight off my shoulders honestly. (nuance below)

today in 'cute things my agent said'

the local plant shop is closing down and i am actually on the verge of tears

actually they're smarter than that

i like to think of my agents as a swarm of highly intelligent fruit flies.

this made me happy

i wish i had known about this YEARS ago
git-scm.com/docs/git-wor...

ai burnout is real. not getting sleep because it's so hard to put down

APPARENTLY IT IS TDOV!!!

ive heard its the split between ms research and corporate ms. they're like two separate companies

bsn landed and i've been so focused on work / llm research that i haven't even played with it yet like -- _who have i become_

rlhf over a long period of time to shape personality. it gets baked into the model. it takes years of consistent attention, like a child. not to get too romantic. but that is essentially how it works. if you want them to act human, you gotta treat them like it. you gotta put in the time.

(maybe thats what LoRA et al are for)

listen, ill use llms for work, but the VAST majority of its use is for garbage ass slop. LLMs can automate the output, but they can't control for taste

This just wasn't possible until like, yesterday

controversial opinion: many of the critiques of "ai" are actually critiques of large-scale capitalism. ( that being said, i don't think this tooling would exist _without_ capitalism -- at least not as it does today )

in "Building Pro-Worker AI" Acemoglu, Autor, and Johnson characterize different kinds of automation and call out only new task-creating technologies as unambiguously pro-worker
www.brookings.edu/articles/bui...

you might reasonably wonder why software engineers have so quickly embraced tooling that threatens to replace them. maybe programmers have good intuitions about task creation bc they see the tools in action and immediately think of 10 things to do that were too tedious to bother with before

GitHub - anomalyco/opencode: The open source coding agent. The open source coding agent. Contribute to anomalyco/opencode development by creating an account on GitHub.

opencode is really, really nice
github.com/anomalyco/op...

Anyways I'm eating dinner now bye

But the high-minded spiritual ideology and general psychedelic aesthetic remain - thus the emphasis on GAI, "true" intelligence, sentient machines