still low on energy these days, so I should acknowledge that I am probably not supposed to feel like a museum piece by the comment about Before Times... but I don't remember ever having a thought in the shape of "this app should exist" myself, so yeah, I probably do feel like a museum piece now

as for the more-likely-intended genuine interest about my closest examples I can think of how I deal with these kinds of situations:

• if I spot a bug in software I am using and I have energy to report it with repro steps, I try to do that - not exactly asking a "buddy" but whatever, e.g. https://github.com/microsoft/TypeScript/issues/41707
• if I randomly visit a friend and they are cooking dinner and they ask me if I want some, I say yes (..as a stereotypical cis man when it comes to free food)
• when I visit a website a lot and their dark mode has big eye-piercing white button, I go ahead and add a style for myself using https://addons.mozilla.org/en-US/firefox/addon/styl-us/ to make that button dark:

• when I wanted to see a spreadsheet of buildings in a game, I made one .. and why not open source it too even if probably no one else will ever use it .. https://peter.hozak.info/urbek/#?c0=Desert,c37=\+,
  • or long time ago when I joined https://github.com/odegroot/Anno-2070-data-extraction so that I could make the optimized layout shown as the first table row on https://anno2070.fandom.com/wiki/Eco_%26_Tycoon_Housing_Layouts#Without_Monument
• in a recent PR, in the code slop from @mruwnik, I can see what was the tradeoff that the imperative DOM shit was simpler for the team at the time to do and that the code could have been improved "later" if needed for more complex features .. and now @Olivier Coutu with the help of coding assistants will need to parse my comment about potential memory leaks - is that situation better with coding assistants or would it have been better if the current team was forced to pay some human developer if they wanted new features to be developed? (although the "average" humans would probably produce slop code too)
• if I have a dispute with my husband, we argue in person
• if I have a dispute with my ex, we tend to have difficult conversations over whatsapp
  • is whatsapp a dispute app for me? no.
• if I am interested to continue a conversation about how exactly we seem to misunderstand each other in a comment section of a lesswrong post, I write one more comment
  • is lesswrong a dispute app for me? maybe.
• if someone could benefit from a vibe-coded dispute app even though I am clearly not the target audience myself, does that sound like a good thing to happen in the world?
  • yes
• is vibecoding going to be a net benefit for the vibecoders and for humanity?
  • I've set up a calendar reminder for 1 year from now to ask @sarahconstantin that here .. no idea myself, just a big fat bias towards "oh god no"

Is this wish compatible with not throwing away a free lunch?

Have we became so anti-social that the only 2 options are to do it alone or not at all?

I'm afraid that I do understand your point of view - I feel myself very exhausted for the last few years so I was not helping my friends in open source lately, so they opted for coding assistants instead and now when I see the code I feel recoil from the AI slop and I do not wish to return to the project. If they want things done and I don't "want" to help, what are their options?

Brave new world we live in, infinite productivity increase from zero to something for people who don't have time to became good at a craft, burnout for a few of us who used to be good and well paid but became overwhelmed by the ever-ready Waluiging incompetent assistant attractor.

eg when the whole point of function A is to call function B under certain conditions, Claude may just…forget to call function B. and not fix this, after repeated reminders.

aaaah 😱 how are there people who don't find this completely utterly insane to accept such a behaviour from a coding tool? 

for me, it's like an elevator that "sometimes" jumped half a meter and then refused to go to some floors - I would call the emergency repair line if that happened and not try to excuse it that "it's so much more convenient than the stairs, even if you have to press the 6th floor button multiple times - it might drive you to the 12th floor first, 4th floor second, but it will almost certainly work on the 3rd try" ... and if I broke my leg (~didn't know how to program in some language), this unreliable elevator would sound MORE scary to me, not less

I think I must be missing some kind of adrenaline enthusiasm that makes me less excited around hype for an incompetent technology that will probably kill us all not long after it gets actually competent ... or just generally becoming a grumpy old man.

It’s not that it weakens your point, but it’s that starting a sentence with “It’s not that …” triggers audiences to narrate your writing in “AI voice”. It disintegrates a reader’s brain because just the smell of AI is a noxious fume. Too often it’s a sign of other lurking deficiencies; there’s never just one cockroach in the kitchen.

oh god, I really really hate the self-illustration 😱 enough to think that it's brilliant? not sure yet...

80-90% are falling behind what exactly, please? to not want to decrease your productivity by 20% and leak customer data sounds like surprisingly rational collective behaviour to me.. probably best to pay for chatbot/coding assistant subscriptions to any employee who wants it since they will use it anyway and "free" tiers are paid by data and integrating any "AI" used to attract investors in the last few years, but do you have statistics that paying customers actually want those AI powered products at nondumping prices? did anyone show any non-self-reported measured increase in productivity (in terms of what the company produces for which their customers pay, not lines of code)? did any early AI-first company other than nvidia report profit numbers instead of just revenue? do early adopters from 5 years ago do better than late adopters from 5 months ago?

tbh, "wait until it starts working" might be a good strategy if there is very little first-mover advantage.. AGI is not here yet, not sure any company can prepare for it by adopting current LLMs "more"

Sounds to me like we always have to calculate a social path integral to a level of approximation appropriate to the situation, even in ask culture... If a friend is lactose intolerant and they know I know that, then even in ask culture it would be weird for me to ask if they want some non-vegan ice cream (and they might assume that if I asked, I would be either joking or offering vegan ice cream, not that I was actively stupid) - so I don't see the option for 0 echos tbh, just an option to agree that coarse approximation of social consequences is totally fine in most situations and as a default and that mistakes are better on the side of oversimplification rather than overthinking it and not interacting at all.

Or some questions like "May I cut your wrists?" seem like they are almost never appropriate, perhaps as a joke between the right kind of people, or meta level sarcasm when judging how much someone is genuinely into the ask culture thingy.. number of echoes can be a fraction sometimes..

So I would imagine that not seeing public comments as needing more social consideration for more diverse audience than DMs is a mistake even in ask culture worth pointing out to people when it could have been formulated with a better escape hatch..

“Adding Error Bars to Evals”

something went wrong with the link rendering => https://arxiv.org/pdf/2411.00640 

I also wish there was no industry that would serve as an example for that employment model...

nah 🙈, the stupid companies will self-select out of the job market for not-burned-out good programmers, and the good companies will do something like "product engineering" when product managers and designers will make their own PoCs to validate with stakeholders before/without endless specifications handed over to engineers in the first iteration, and then the programming roles will focus on building production quality solutions and maybe a QA renaissance will happen to write useful regression tests when domain experts can use coding assistants to automate boring stuff and focus on domain expertise/making decisions instead of programmers trying to guess the indent behind a written specification twice (for the code and for the test.. or once when it's the same person/LLM writing both, which is a recipe for useless tests IMHO)

(..not making a prediction here, more like a wish TBH)