Journal Entry For 2026-04-08

• This journal entry is somewhat older - I started it maybe two weeks ago now? So not all things are as "fresh" as they maybe could be. Regardless, I wanted to get this out.
• I have so many conflicting feelings about this whole "AI" thing. I have a negative reaction to what I perceive as purist arguments but I also am an "Incrementalist": I do think change is possible, but very much not in the way some people seem to think (i.e. by abstaining). As I said in my critique of critique article: "I think Doctorow, Tante and me we all oppose purism, we just all diverge on what follows from this: Doctorow seems to think reappropriating technology is enough, Tante believes in critique and I don't believe in either. I think we are shaped more by society and its structures, the systems that shaped and still shape us. If we are able to change anything it's incremental, minuscule and less grandiose than what these thinkers seem to believe."
• That said, I am kind of happy that we are hitting the usage limits of the (totally subsidized) Codex plans at work, which means we will have to actually interact with the real limit of using "AI" (if not the real costs, yet).^[1]
  • This means using "AI" more intentionally and will hopefully curb the frivolous use of "AI" coding tools. I think "AI" use coupled to reality is a totally different beast than "AI" use decoupled from it.
  • This also hopefully means people will stop outsourcing the thinking part of working with "AI" coding tools.
  • UPDATE: I wrote this like 10 days ago. The company has since lifted the token limit instead of maybe at least building more token consciously.
• I finally started to continue listening to my audio book, which I count as a big win as I didn't get any book reading/listening in in weeks.
• And an even bigger win: I went on two runs in the last few days after not running since October last year! I restarted (for the third time now) my Watch To 5K plan - maybe I'll stick with it this time. I almost made it last year (I think 4 runs were missing out of the 36) - with lots of breaks in between, as I had needed a surprising amount of recovery time and also that life happened made it hard to get back into it, when I was finally recovered.
• (This is a privileged position, I think. Even though I count myself as a normie (with an ok-paying tech job), I do concede that I have a relatively good work computer that I am allowed to use in my free time and on top of that I had some money saved (and could save for such a thing) that I invested in a NAS recently so my tech needs are met at the moment). In a weird way I like the idea of being hardware constrained and that I will probably have to make due with the hardware I have right now for a while. If I buy anything, it will be expensive, so I will have to make very sure that I actually need it. I hope this makes me interested in more efficient approaches to using and keeping data. That I will prefer less resource hungry software and so on.
• I love this persons writing so much, so here's a longer quote: "To repeat it loud enough so those in the back can hear, I’m not suggesting generative AI is valueless. If they can solve both the economic and legal/moral issues surrounding it—and those are, to be clear, very fucking big ifs—it has a bright future in verticals where it can do what it’s best at: tightly constrained, non-creative generation. As a former programmer who’s seen what a lot of commercial code actually looks like, I assure you that you and I both, right now, rely on 100% human-written code shittier than what Claude Code farts out. Putting unreviewed LLM-generated code into production is insane, but as long as there are humans who read, understand, and verify the generated code, it’s going to be fine. […] But beyond code generation, we fall off the cliff of diminishing returns real fast. Chatbots are good as a jumping-off point for web searches and research, but you can’t rely on them. They’re not good at writing text that requires, or is even just improved by, any kind of verve or voice or original thinking—so that leaves, what, first-level customer support responses? Meeting summaries? Business memos? I’m aware of, and sympathize with, the “this way lies Idiocracy” concerns here, but I suspect they’re going to prove overblown. [… W]e need to have another, less tech-focused and more fundamental revolution first. As much as possible, we need to choose smaller businesses to both buy from and work for; when we choose to buy from or work for bigger companies, we need to take into account what they’re doing for, or to, our environment, our privacy, and our politics." Another great essay by Watts Martin.

• ← Previous
  DailyDogo 1602 🐶