I’ve been talking about this a lot with founder/coder types in my circle of friends with a wide variety of opinions.
My theory as an old guy is that the standards will just go up.
There’s a current business model where you can make a basic but useful tool that solves a specific business problem and make money. That’s going to end.
We’ve seen this before. A good example would be when the mobile app stores launched and you could get traction with just about anything. And then you couldn’t.
> There’s a current business model where you can make a basic but useful tool that solves a specific business problem and make money. That’s going to end.
I don't know... Because the tool that solves a specific business problem usually requires tons of business expertise. And when company buy this tool, they mainly do it for the expertise diluted in it.
If they didn't already made their own in-house implementation, it's because they don't want to invest in maintaining the tool that requires expertise outside of their actual business.
Meanwhile, the company building the tool can invest in keeping this expertise because it's financed by the multiple companies paying for the tool.
That's what I think will happen. Not just standards but taste and design as well. I think there's almost even more demand for good designers now than ever before.
Why? VERY good design signals this is tasteful and quality. Not an AI-slop-vibe-mess.
> Why? VERY good design signals this is tasteful and quality. Not an AI-slop-vibe-mess.
I'm not too sure about this. Very good design, at least in terms of marketing materials such as screenshots, may be end up signalling the use of AI in the coding as well. I know if an app has obviously-AI-generated photos in the screenshots I would be disinclined even to try it.
It doesn't seem obvious that it's a problem for LLM coders to write their own tests (if we assume that their coding/testing abilities are up to snuff), given human coders do so routinely.
This thread is talking about vibe coding, not LLM-assisted human coding.
The defining feature of vibe coding is that the human prompter doesn't know or care what the actual code looks like. They don't even try to understand it.
You might instruct the LLM to add test cases, and even tell it what behavior to test. And it will very likely add something that passes, but you have to take the LLM's word that it properly tests what you want it to.
The issue I have with using LLM's is the test code review. Often the LLM will make a 30 or 40 line change to the application code. I can easily review and comprehend this. Then I have to look at the 400 lines of generated test code. While it may be easy to understand there's a lot of it. Go through this cycle several times a day and I'm not convinced I'm doing a good review of the test code do to mental fatigue, who knows what I may be missing in the tests six hours into the work day?
> This thread is talking about vibe coding, not LLM-assisted human coding.
I was writing about vibe-coding. It seems these guys are vibe-coding (https://factory.strongdm.ai/) and their LLM coders write the tests.
I've seen this in action, though to dubious results: the coding (sub)agent writes tests, runs them (they fail), writes the implementation, runs tests (repeat this step and last until tests pass), then says it's done. Next, the reviewer agent looks at everything and says "this is bad and stupid and won't work, fix all of these things", and the coding agent tries again with the reviewer's feedback in mind.
Models are getting good enough that this seems to "compound correctness", per the post I linked. It is reasonable to think this is going somewhere. The hard parts seem to be specification and creativity.
Maybe it’s just the people I’m around but assuming you write good tests is a big assumption. It’s very easy to just test what you know works. It’s the human version of context collapse, becoming myopic around just what you’re doing in the moment, so I’d expect LLMs to suffer from it as well.
> the human version of context collapse, becoming myopic around just what you’re doing in the moment
The setups I've seen use subagents to handle coding and review, separately from each other and from the "parent" agent which is tasked with implementing the thing. The parent agent just hands a task off to a coding agent whose only purpose is to do the task, the review agent reviews and goes back and forth with the coding agent until the review agent is satisfied. Coding agents don't seem likely to suffer from this particular failure mode.
Cowardice is the wrong word here. It implies that these people have a desired action that they're not taking because of fear, weakness, or hesitation.
What tech companies actually have is rapacious sociopaths for leaders. They have purposely brought about the current state of affairs through intensive lobbying, spending, and direct action.
For the most part, they don't believe that they should be held accountable for their behavior. They don't fundamentally believe in democracy, and many of them don't really believe humans and human life are more important than some other abstract concept that they have in their heads. At root, they all believe in rule by the elite.
This may seem like an argumentative distinction, but I would counter that it's crucial to understanding what we have to do next, which is not to try to convince them, but rather to take back the power that they've accumulated over us, against their best efforts to stop us.
> take back the power that they've accumulated over us, against their best efforts to stop us.
It sure would be nice if so many smart technologists hadn't spent the last two decades poo-pooing distributed p2p software in favor of centralizing https and javascript webcrapps because "they have nothing to hide" and can always "vote with their wallet".
> basically unreachable by public transport, or bus that goes 2x a day on some days of the week.
This sentence is hilarious from an American perspective. There are central business districts of major US cities that are less connected to public transit than the most remote rock at the end of a steep canyon in Switzerland.
A bus that ran 1x a day on any day of any week would be a drastic improvement for nearly all of the US.
Its not mutually exclusive - most of the world thats not in stone age has better public transport than US, I guess everybody knows that and its not by accident but for good (well bad but logical) reasons.
That some PT is still not covering somebody's full needs for long term living is understandable too I presume, especially if its few days gaps in service.
Curious what the point you're making here is. I don't know anything at all about Bob Park and whether he is a crank. But if you make your career doing the admirable work of debunking pseudo-science and nonsense theories, you would necessarily be linked to in discussions of those theories very, very frequently.
So maybe that's not a good description of him. But the link you posted is hardly dispositive.
The pseudo science of corporate values is a contradiction in terms invented by HR ladies who drink tea for a living. People who believe such things also believe in aliens and have theories about vegetarian tigers.
You are now debunked.
(This comment is intentionally stupid, useless and the author knows nothing about the topic)
What I'm talking about is orgs like the CDC, FCC, DMV, IRS, etc. They grow more than they re-org and cut the workforce, moreover, they seldom redefine positions to take on more responsibility at the staff level. Obviously they don't have the same outsourcing threat due to some limitations --which is good, but it's also true that public companies can outsource more readily. Goverment workers rarely get fired for under-performance outside of budget-induced cuts.
reply