I find interesting the systemic explanation of Bruce Bueno de Mesquita and Alastair Smith in "The Dictator's Handbook"^1: In any publicly traded company, if the executives (or the board) are not ready to do whatever it takes to maximize profit, they will be replaced by people who are. It becomes a selection process creating tyrants. If you're lucky (as employee, customer, or human living on the same planet), whatever it takes might be aligned with employees and customers' interest. When times are bad, whatever it takes has no limits, it becomes a question of survival or progression for business leaders.
I'm curious to see how much that maps with what you identified in your new book! Patagonia is private and under the control of a few benevolent dictators ; Costco and Nordisk are a bit more surprising, I'm keen to know more.
The fundamental question is what the people who work at an org are incentivized to maximize for "whatever it takes." Today, by default, orgs are aligned to maximize shareholder value (even at the expense of long-term viability) but there is nothing natural or logical about this.
I love mountain biking, and part of it is the effort it requires to get to the high trails, fitness to do multiple laps, and the quiet of being in the forest.
But for the past 5-10y eMTBs became a thing and progressively became more affordable. And now you've got 200lbs gorillas zooming past you, digging into the trails with their heavy rigs, and a large number of people who'd never have paid for the effort of getting there without their ebikes.
> Users don’t care whether the code was written by AI or by hand
That's true, but they care deeply about the consequences of that:
> about how AI “writes bad code,” “introduces bugs,” “creates technical debt,” or something along those lines.
So whomever your strawman is, they got a point.
Note that I'm "anti-ai", I use it a fair bit and even received the trendy email asking me to watch out how much I spend in it cause it's expensive. I'm also not delusioned into believing the "it's 10x faster" and "code doesn't matter anymore" marketing. If the thing fails it's my name on the git blame and my number they call at night so I'll review that code thank you very much.
I feel like past the wow effect it's pretty easy to see the seams and the limits, even on "frontier" (god do I hate that term) models, and nothing replaces human skill for now if you're working on something with any significance.
Dang sums it all, I dont perceive hn as being pro or against AI, it's a mix, but if you're polarized, whatever "side you're on" you'll feel the other side is over represented.
> Dang sums it all, I dont perceive hn as being pro or against AI, it's a mix, but if you're polarized, whatever "side you're on" you'll feel the other side is over represented.
It doesn’t feel that way to me. I remember reading the recent thread about Bun being rewritten in Rust with LLMs. Definitely not a 50/50 split of positive and negative comments. It was a pile-on.
I actually see surprisingly little "anti AI" in that thread at all, and mostly just concerns about merging 1,000,000 lines of code with minimal review, which i think is a reasonable thing to be concerned about even if you use LLMs yourself and aren't "anti AI"
I remember a moment at the very onset when chatgpt was just opened to the public. A manager sent us a congratulations text chatGPT had prepared for his kid's graduation from high school. He said that he was not much of a writer and that was miles better than he could have said. We had a discussion about the moral side of that (though moral might not be the word), pointing out that bad words that come from a genuine place and required effort from you are better than great words that are manufactured by a robot. He didn't see the issue.
I think this just depends on a system of values, "to each their own". I don't see the point of having a bot write comments for me on HN, or blog posts for me, or answers on GitHub. I feel great for articulating my thoughts in a way that (narcisticaly) I can enjoy re-reading myself. Some people don't value that, and for whatever motivation don't mind delegating their voice to a bot.
And then there are the "people" who just try to build accounts with lots of internet points that they might be able to resell for a few bucks. Those can die.
> “If you don’t learn how to use AI, you’re going to be left behind.”
I don't subscribe to that even to begin with. Learning "how to use AI" in a a dev workflow takes less than a month, just by practice. Assuming "how to use AI" refers to using agents, rather than just a more advanced auto complete, once you've made the couple of mistakes you can make a couple of times (mainly, leave the coding agent to do something too big for too long), you're caught up. Maybe it take a little more time to get the habits into your workflow.
Leveraging a LLM for coding is orders of magnitude simpler than the code it's writing, and the skills you need to review said code.
Sounds like an interesting topic!
> Why do good companies go bad
I find interesting the systemic explanation of Bruce Bueno de Mesquita and Alastair Smith in "The Dictator's Handbook"^1: In any publicly traded company, if the executives (or the board) are not ready to do whatever it takes to maximize profit, they will be replaced by people who are. It becomes a selection process creating tyrants. If you're lucky (as employee, customer, or human living on the same planet), whatever it takes might be aligned with employees and customers' interest. When times are bad, whatever it takes has no limits, it becomes a question of survival or progression for business leaders.
I'm curious to see how much that maps with what you identified in your new book! Patagonia is private and under the control of a few benevolent dictators ; Costco and Nordisk are a bit more surprising, I'm keen to know more.
^1 https://en.wikipedia.org/wiki/The_Dictator%27s_Handbook
reply