“So I guess what I’m trying to say is, the new workday should be three to four hours.“
-
“So I guess what I’m trying to say is, the new workday should be three to four hours.“
Yup! That's what every worker knows and should have been fighting for with solidarity for decades. Every neurodivergent person knows that we can't do concentrated work for more than 3 hours, and that extended hyperfocus blocks drain our energy for the next day. It's not sustainable.
Steve Yegge writes about how AI + Capitalism creates an energy vampire https://steve-yegge.medium.com/the-ai-vampire-eda6e4f07163
@saraislet I 100% agree... except this goes for any "speed improvement" tech. Although I'd say that AI has only slowed me down because I get to now review the shit people have been able to do own their own but now instead "vibe" their way through it with crappy results that I need to fix ^^'
I'm a happy non-user
-
“So I guess what I’m trying to say is, the new workday should be three to four hours.“
Yup! That's what every worker knows and should have been fighting for with solidarity for decades. Every neurodivergent person knows that we can't do concentrated work for more than 3 hours, and that extended hyperfocus blocks drain our energy for the next day. It's not sustainable.
Steve Yegge writes about how AI + Capitalism creates an energy vampire https://steve-yegge.medium.com/the-ai-vampire-eda6e4f07163
@saraislet know, sure. but not accept. it tore me apart and i’m still mad at myself that i only went back to the grind part time (cue brain worm saying that time spent writing this is stolen from $employer)
-
@saraislet ...did he write something sane this time? The last thing of his that I read was, uh, not good
@BoredomFestival this has at least some good points and some
🤨
🫣🫠 -
@saraislet know, sure. but not accept. it tore me apart and i’m still mad at myself that i only went back to the grind part time (cue brain worm saying that time spent writing this is stolen from $employer)
@doomsey what don't you accept about it?
-
@BoredomFestival this has at least some good points and some
🤨🫣🫠@saraislet OK, this one is actually readable and does have some interesting points, but the elephant in the room is that he asserts that it will 10x your productivity, but it utterly exhausts you to do so... without explaining why it's so much more exhausting than "ordinary" work.
-
@doomsey what don't you accept about it?
@saraislet Ah, that’s the problem. There’s a big difference between knowing something is true and having that truth be something you can feel. I’ve known that being productive a few hours in a day was all I could manage for a really long time. But I didn’t ever accept it, at best coming to a detente with myself of sorts that blank-stare time was work.
And then I ended up in a situation where I was told 40 hours is mandatory, and I fell apart.
-
@saraislet OK, this one is actually readable and does have some interesting points, but the elephant in the room is that he asserts that it will 10x your productivity, but it utterly exhausts you to do so... without explaining why it's so much more exhausting than "ordinary" work.
@BoredomFestival strong agree — if it exhausts you, then it isn't sustainable. That's why I think it's akin to trying to maintain hyperfocus 8 hours a day, 5 days a week, consistently. It's just not feasible for humans in the long-term.
-
“So I guess what I’m trying to say is, the new workday should be three to four hours.“
Yup! That's what every worker knows and should have been fighting for with solidarity for decades. Every neurodivergent person knows that we can't do concentrated work for more than 3 hours, and that extended hyperfocus blocks drain our energy for the next day. It's not sustainable.
Steve Yegge writes about how AI + Capitalism creates an energy vampire https://steve-yegge.medium.com/the-ai-vampire-eda6e4f07163
This is an honest post about some of the practical challenges of being a software engineer using AI.
Siddhant shares concrete relatable examples of many common experiences among software engineers, from context switching and code review fatigue to tool churn and engineering perfectionism clashing with nondeterministic AI output.
Siddhant also suggests specific practical changes to help software engineers struggling with the daily habits around coding with AI. (These are probably most useful to early-mid career folks as they're lessons likely already learned by more senior folks, but could be worth sharing with colleagues!)
https://siddhantkhare.com/writing/ai-fatigue-is-real -
This is an honest post about some of the practical challenges of being a software engineer using AI.
Siddhant shares concrete relatable examples of many common experiences among software engineers, from context switching and code review fatigue to tool churn and engineering perfectionism clashing with nondeterministic AI output.
Siddhant also suggests specific practical changes to help software engineers struggling with the daily habits around coding with AI. (These are probably most useful to early-mid career folks as they're lessons likely already learned by more senior folks, but could be worth sharing with colleagues!)
https://siddhantkhare.com/writing/ai-fatigue-is-realThis is a clear, short explanation of some of the dynamics underlying software delivery velocity, and what does vs doesn't change with AI
While it's debatable exactly what or how much we need to understand about a system (when no single person can comprehend the entirety of modern complex systems) — we need to comprehend *something* about the manifested behavior of the software we deliver (with the security capabilities and infrastructure platform capabilities being nontrivial aspects of this).
Jesse Landry argues rightly that when velocity outpaces comprehension, fragility compounds on the gap between knowledge and reality. In other words, another form of tech debt.
https://www.devcuration.com/the-velocity-trap/ -
This is a clear, short explanation of some of the dynamics underlying software delivery velocity, and what does vs doesn't change with AI
While it's debatable exactly what or how much we need to understand about a system (when no single person can comprehend the entirety of modern complex systems) — we need to comprehend *something* about the manifested behavior of the software we deliver (with the security capabilities and infrastructure platform capabilities being nontrivial aspects of this).
Jesse Landry argues rightly that when velocity outpaces comprehension, fragility compounds on the gap between knowledge and reality. In other words, another form of tech debt.
https://www.devcuration.com/the-velocity-trap/@saraislet or more accurately, Jesse Landry's LLM argues.... it cracks me up, basically 100% of the opinion pieces about AI - pro and con! - in my feed are substantially written by AI
-
R relay@relay.infosec.exchange shared this topic
-
@saraislet OK, this one is actually readable and does have some interesting points, but the elephant in the room is that he asserts that it will 10x your productivity, but it utterly exhausts you to do so... without explaining why it's so much more exhausting than "ordinary" work.
@BoredomFestival @saraislet one way I read this is that using these tools changed their emotional state (to "hyped") and therefore they are just working more and more consistently.
-
@BoredomFestival @saraislet one way I read this is that using these tools changed their emotional state (to "hyped") and therefore they are just working more and more consistently.
@BoredomFestival @saraislet a.k.a., placebo
-
This is an honest post about some of the practical challenges of being a software engineer using AI.
Siddhant shares concrete relatable examples of many common experiences among software engineers, from context switching and code review fatigue to tool churn and engineering perfectionism clashing with nondeterministic AI output.
Siddhant also suggests specific practical changes to help software engineers struggling with the daily habits around coding with AI. (These are probably most useful to early-mid career folks as they're lessons likely already learned by more senior folks, but could be worth sharing with colleagues!)
https://siddhantkhare.com/writing/ai-fatigue-is-real@saraislet "I had a prompt that worked perfectly on Monday. Generated clean, well-structured code for an API endpoint. I used the same prompt on Tuesday for a similar endpoint. The output was structurally different, used a different error handling pattern, and introduced a dependency I didn't ask for.
"Why? No reason. Or rather, no reason I can access. There's no stack trace for "the model decided to go a different direction today." There's no log that says "temperature sampling chose path B instead of path A." It just... happened differently."
This is why it's so difficult to work with AI. Why does it introduce dependencies I didn't ask for? Why does it not do things in a deterministic way? And what's more scary for me is: Why are some others not more worried about this?
