No, and that’s a pretty telling point, if you think that hiring a programmer only gives you someone writing code, the problem solving and complex engagement for feedback is something you get from that
And honestly, based on all the “AI” things I’ve seen so far, and having extensive programming experience, I’m not convinced you have the required knowledge and experience in programming to be able to tell the difference
Would love to check your CV, but don’t think that starting an experience competition would yield to much.
Have it your way and I’ll do what I do.
I’ll take that to mean that you’re not experienced in it, that’s fine, but I’ll take your evaluation of “AI” tools for these tasks with a grain of salt (you can look at my GitHub profile and my over 500 Godot contributions if you’re interested)
Admitting, you have quite impressive experiance with Godot. I guess you know your C++ stuff.
I can well admit that you probalby beat me there. But still, should I now just close my Github account and do my own programming when I’m happy with what it delivers ?
I don’t think that was the point at all, I genuinely don’t think anyone wants you to stop doing what you’re doing. I think it’s just a matter of seeing things from another perspective and having a discussion about it.
Indeed all I’m saying is that it’s important to consider that if you’re experience with programming is limited your ability to judge the quality of an LLM doing programming is also limited
–plus, (apologies for jumping in randomly)
as @dragonforge-dev stated, that vibecoding will be a “richman’s” operation in time,
it is good to be able to have a second route so it will not end your programming career. (its good to know your stuff is all, otherwise, do what you feel best)
It’s not like I don’t have any programming experience. I’ve coded 6510 assembler back in the 80s.
I’ve coded C++ for large banks here in Switzerland that handled Billions of $$$ on a monthly basis.
Currently I’m mostly focusing on getting features in. I’ll check the code quality later.
So far I can tell that the simulator runs smoothly. No microstuteres or anything like that with a really large scenery aera. So my requirements are met from that perspective. I’m sure that when I look into the code itself that I will find manythings worth optimizing though.
You do have a point which is buried here I think, and is important. Which is, you are investing money for an end goal. It makes sense.
I get that giving up streaming subscriptions is a way to afford those things. But I, like many people, share those subscriptions with others and therefore the idea of giving them up means that 1) It impacts multiple people in my family, etc. 2) I don’t actually get all that money to spend somewhere else because it’s not all my money. To @athousandships point, not all people have that disposable income.
And if you are comparing it to hiring a developer, well yes, most people would never even consider hiring a developer. As a developer who people hire, that amount of money per year obviously wouldn’t allow me to avoid starving as a homeless person, let alone be tenable for me to make a living. But employing a person is a different scale of economics completely.
Are you saying that you haven’t even validated the code? And yet you say:
How do you know that if you haven’t even looked at the code?
roughly 10 years ago I was part of a group of people wanting to build a flight simulator. I mean a real cockpit with motion etc. We laked some needed softare so, some one with no programming experience at all suggested seriously to hire a programming team from India. He believe they could write the software completele for $10k. I only asked him 2-3 questions on how he would do quality checking and stuff like that to make him see that it would not be that simple to get this software we needed written.
I have looked at some of the code. But only cursory. It’s not my main focus right now.
I see the results. It’s not pacman. It’s a functioning flight simulator that contains the features I asked it to program. If the LLM used auto wrong or missed some const optimization is not relevant for me at this stage.
Then I’d suggest reserving judgement on it and comparing it to a real programmer until you have, otherwise I’d say that’s pretty dishonest
Have you ever shown someone work in progress and they shaked their head about it ?
What does that have to do with anything?
I’d rather say it sounds like standing next to a bridge under construction and saying “oh I have confidence in the skill and limitations of the people hired to build this bridge, sure I haven’t looked under the tarp to see if it’s built from concrete or straw but we’ll work out those details later”
My point and issue is that it’s fine if you want to go about development this way, but if you’re going to argue the capabilities and usefulness of those tools, and comment on how good they are, you might want to actually look a bit more closer than you might if you’re just using them for yourself
Well I have enough experience in software development to know that there is nothing more permanent then a temporaty solution.
By the way it’s not that long ago that a bridge in Genova colapsed. Big time.
Nvidia came to a different conclusion:
‘The cost of compute is far beyond the costs of the employees’
Many more are going to find that out soon enough I think, if they don’t watch out
If you read the details of the aricle it’s rather telling though:
“As a result, some firms are beginning to reevaluate AI not as a clear cost-saving substitute for labor, but as a complementary tool—at least until the cost structure stabilizes,” he said.
…
“It’s not just about AI becoming cheaper than humans,” Lee said. “It’s about becoming both cheaper and more predictable at scale.”
So in the end it’s mostly about calming people that fear for their jobs. I didn’t see mentioned in the article that software developers salaries have droped significantly the last two years, as I’ve read other places.
Good catch. Got this article from this video, that also tries to calm down developers.
