I think it's because not many know how to measure it properly.
I can output 5 useless/bad features in a day with Claude or I can output 1 useful feature per 2 day period. Which one has better impact on ROI?
In this example, it might seem like it's an easy answer. But, in the real world, it is a lot more nuanced and much more difficult to measure and so not many are bothering to do it and are opting in for the simple solution of following the hype.
> planning every last detail before writing code is boring
Not only that but you can't really plan everything. It is impossible. Without LLMs, with every line of code you are making a decision or discovering something new that must be dealt with or realizing how the current thing might impact something else and so on.
There is no way for a programmer to consider all of these little things ahead of time and if an attempt is made, it will take as long as actually writing that code.
> Without LLMs, with every line of code you are making a decision or discovering something new that must be dealt with or realizing how the current thing might impact something else and so on.
Part of this is true, part of it the agents catch at least a non-trivial portion of. If you prompt it to do a review, especially with a specific angle like ensuring sustained write performance, or how it will work when the future extensions are implemented, they do often catch a lot of issues.
I agree you lose a fair bit of the sense of "it feels like I'm doing something wrong", or "this doesn't seem optimal" etc. I think the skill in using these tools is to determine when you need that control and where it doesn't really matter.
Reason dictates having redundancy in place. Having prepared scenarios for what to do. A lot of countries clearly don't have that and they are operating on the assumption that no major disruption is going to happen.
Depending on resources coming from historically unstable locations and not having plans to prepare for such instability is just foolish.
I recently built a home server and used Noctua fans for the first time. I absolutely love how quiet they are. If I didn't know the room had my server in it and was running, I wouldn't even notice the very quiet sound of the fans.
I am running them at about 800rpm and the CPU is usually between 33~37 degrees.
When I rebuild my main PC, I will surely go with them again.
For me, almost every single time a conversation like this happens in real life it boils down to the one side claiming that "This is the future" and "Don't get left behind" followed by a torrent of hype and buzzwords. So no, there is no interesting conversations to be had about LLM programming anymore.
Some interesting conversation one can have with coworkers specifically:
1. How should code review and responsibility for code be updated to a) increase velocity, b) keep quality and c) keep reviewers from burnout. There are plenty scenarios in which vibe coding a component in an afternoon is the correct choice, even if it is buggy, insecure, and no one really understands it.
2. Which parts of the codebase work well with code assistants, which don't? Why? What could be changed to make it easier? In my experience, Claude Code sometimes loses its mind on infra topics. It is also not very good at complex, interconnected services (humans aren't either).
3. Which tasks could be offloaded to agents to save everyone time and sanity? - Creatig Jira Tickets from meeting transcripts is an obvious one, collecting and curating bug reports another one.
4. How should we design systems to better work for coding agents? Does it influence our tech choices? Should it influence them?
5. Is AI a net positive or negative for security?
And so much more. The last topic in particular is incredibly important, and things are developing so fast that you can probably have a new conversation on it every two weeks.
Maybe you struggle to have good conversations because I just provided an anecdote and you immediately stated that my anecdote is false? If this is how you typically interact with people I’m not surprised you’re not having interesting conversations.
He didn't state your anecdote is false, the first two words of his comment are "for me". That means in his experience, not yours.
Ironically, in your crusade to wave the "I'm being censored!" flag, the only person who is trying to do any censoring... is you!
And to top it off, as if that wasn't enough, you're also incredibly snarky and basically implying that this person who just vaguely disagrees with you must be unlikable or something. Which, in another twist of irony, actually makes YOU appear unlikable, because what well-adjusted adult would feel the need to throw someone under the bus for slightly disagreeing with them?
The problem is that mrcsharp added the last sentence: "So no, there is no interesting conversations to be had about LLM programming anymore." So they are definitely trying to turn their anecdote into a universal.
Most of your comments about johnfn are still apropos, though...
> "there is no interesting conversations to be had about LLM programming anymore"
The first sentence was in his experience, but this is a universal assertion. He is claiming no one, in the world, is having interesting conversations about LLMs.
Is my response really so off-base? Imagine you say "I like Rust because it made my app go fast" and someone replies "There is no one who has used Rust to improve performance." Do you really think that's a normal way to respond to someone sharing an anecdote?
But that's not how he responded because there's a whole ass comment before that.
Okay sure, if you read the comment and then use a Men in Black mind wiper thingy before reading the last sentence, then it might seem brazen or universal. But that's not what you did.
The only way that last comment can reasonably be taken to mean "for everyone on Earth" is if you did not read the lines before it. Because, in that context, to me, it's clear he is only talking about his experience.
This is a phenomena I've noticed lately where everyone feels the need to add a disclaimer for everything and not doing so is seen as an "aha gotcha!" type thing. But we're not algorithms. You do not read one line at a time and then digest it.
You're human, he's human, and there's context. You know that it would be extremely unreasonable for someone to think that nobody, anywhere, has anything to say about LLMs right? Okay. That doesn't mean that this person is being unreasonable.
Read his comment in context. The comment thread, condensed, is:
mudkipdev says "There can't be any interesting discussion about AI programming". We agree that is a universal claim. (Right? I mean, your comment says "You know that it would be extremely unreasonable for someone to think that nobody, anywhere, has anything to say about LLMs right" but isn't this a clear example?)
I say "There can be". We agree that is anecdotal.
mrcsharp says "For me, [stuff that supports mudkipdev]. Therefore, there is no interesting discussion".
He is re-asserting mudkipdev's point. mudkipdev says A, I say !A, he says, actually, A.
Your interpretation has him read the back-and-forth between me and mudkipdev, and respond to "A", "!A" with "B". If you only read mrcsharp and nothing else in the thread I can understand this reading, but the context changes things.
English is my 3rd language. I still disagree with using an LLM to write on one's behalf. I either get to read your thoughts in your voice or the comment is getting a downvote/flag.
> AI is a great equalizer when it comes to communication in English.
Good argument for it but I think 80/20 split applies here. It is likely that 80% of the time it is used to farm for upvotes and add noise.
> And despite what people say, the way you write is very much judged as an indication of your education and intelligence.
I have come across plenty of content and online interactions in English where English was the Author's 2nd or even 3rd language and I find that putting a small disclaimer about this fact is more than enough to bypass such judgement.
I can output 5 useless/bad features in a day with Claude or I can output 1 useful feature per 2 day period. Which one has better impact on ROI?
In this example, it might seem like it's an easy answer. But, in the real world, it is a lot more nuanced and much more difficult to measure and so not many are bothering to do it and are opting in for the simple solution of following the hype.