Ari
banner
arithmetc.bsky.social
Ari
@arithmetc.bsky.social
Mean to the right people and angry about the right stuff. Ex-software engineer and product designer building Desk Job Design Lab in Brooklyn as a community space for hands-on making in an increasingly synthetic world.
Also his feeds would show him more of what he engages with, so maybe he watches a lot of masculine women and feminine men on Instagram reels
November 26, 2025 at 3:15 PM
Completely agree, and will try to make that more prominent when talking about AI on here
November 26, 2025 at 2:09 PM
Do you feel that it’s good to be able to connect with people on the internet?
November 26, 2025 at 2:08 PM
This is like saying a table saw isn’t useful to you or you cut your finger off because you don’t know how to use it/the tool doesn’t fit your workflow. Have some damn curiosity holy shit.
November 26, 2025 at 2:07 PM
Define AGI
November 26, 2025 at 2:04 PM
That’s still in capitalist framing though. We need to change who owns these tools and who benefits from them, not try to wish or moralize them out of existence. Most software engineering jobs are stupid because they are building stupid things! I guess I just think we should think bigger.
November 26, 2025 at 1:59 PM
I certainly think AGI is stupid and any attempt to invent Machine God should be treated as dangerous/dumb and laughed out of the room along with humanoid robotics (aspiring slave masters) and AI image/video generation (waste of resources, like designing by slot machine)
November 26, 2025 at 1:57 PM
If that’s the depth of what you thinking building alongside an LLM is, copying text and generating pictures, then yes, you should stay in your existing workflow because it sounds like it’s working for you! It’s just a tool, and all this moralizing misses the bigger issue: who should own this system?
November 26, 2025 at 1:53 PM
With this level of information literacy, you’ll be sharing screenshots of AI generated IBM corporate SEO slop from the Google Search highlight in no time.
November 26, 2025 at 1:51 PM
And the way I measure my productivity would indicate by that study’s metrics, I am in fact more efficient at building. As a tool, it’s not going away.
November 26, 2025 at 1:50 PM
*next word predictors
November 26, 2025 at 1:47 PM
Yes, I think we should nationalize the tech infrastructure, ban advertising and infinite scrolling feeds, and treat AI (or in this case, next word presenters) as another tool that cannot be controlled by the managers/capitalists
November 26, 2025 at 1:47 PM
It might be very useful! It makes writing code faster! Doesn’t mean we shouldn’t break up or nationalize these tech fiefdoms and tear down the data centers.
November 26, 2025 at 1:45 PM
Eh, evidence would suggest otherwise and it’s odd you immediately went to that as an insult, but I felt it a lot as a female engineer over the last decade so it doesn’t phase me. I’m trying to strengthen AI criticism, not hurt it, by saying our resistance shouldn’t be along this “usefulness” axis.
November 26, 2025 at 1:44 PM
You should when arguing against LLMs and AI. It means the bulk of the argument isn’t on “these things are useless” but “I reject the way they want to restructure society and who they serve”
November 26, 2025 at 1:42 PM
I don’t think Cloudflare or AWS crashed because of LLM usage, but fortunately they do publish post-mortems on their engineering blogs so you can look into it! Not sure how any of this is responsive to: I find LLMs useful for writing software as someone who built pre-LLM for over a decade.
November 26, 2025 at 1:41 PM
I would never tell you an LLM could help you run your farm. It might help you build your farm’s website.
November 26, 2025 at 1:39 PM
Not in my experience. I am telling you: the tool is useful to me. I have over a decade of experience building pre-LLM and ship faster with Cursor in my workflow.
November 26, 2025 at 1:39 PM
No, I’m not.
November 26, 2025 at 1:38 PM
They are a useful tool for me, especially on frontends and dealing with visual design. Sorry that bothers you!
November 26, 2025 at 1:38 PM
If you shrug off those use cases, your argument is infinitely weaker. I’m saying to make a good case against AI, with a theory of power that includes the platform monopolies and effects on labor/the environment, you shouldn’t pretend these tools have NO valid uses.
November 26, 2025 at 1:17 PM
So you don’t use any LLMs as a software engineer? Sorry, my highness. Didn’t realize I was in the presence of a superior intellect! I’ve also built with python and flask (my work is live at the Census Bureau collating HIV/AIDS epidemic data) and I don’t need it explained to me.
November 26, 2025 at 1:16 PM