I think its another example of the Gell-mann amnesia effect, if your'e an expert in something then the AI is often wrong and you're confused why people are saying its great, if you're less skilled then it can be impressive.
That’s a great quote —- when you hire a “creative” to do a job (ie photographer, designer, etc) in area where you aren’t an expert, the conversation of what you want usually starts with similar existing work that you like.
A good creative will take that as a starting point, apply their skills and vision to it, and give you something that solves your specific problem in a unique way, often far better than whatever you had imagined.
In my experience, if you do a similar expertise with AI, it just gives you a facsimile of the inspiration you initially provided and not much more.
> it just gives you a facsimile of the inspiration you initially provided and not much more.
This is not at all my experience but I'm also biased currently making a fair enough living off providing software that does just that, using AI.
In favt I'd say that's the key to enjoyable AI experiences, without strong opinions from the person leveraging, the output is rather bland and corporate.
IBM tried that with CMM (capability maturity model), it didn't work, the problem is NASA knows what they're building, rockets and satellites don't have any grey areas and everything is specified. Other things are less well defined, and the people specifying aren't rocket scientists.
There was a time when the thinking was you can load all the facts into a prolog engine and it would replace experts like doctors and engineers - expert systems, it didn't work. Now its a curiosity
So do you imagine that AI will reach the point that a business guy will say make me a web site to blah blah blah, and the AI will say sure boss and it will appear? Sort of what a dev/team of devs/testers/product managers/BA's would do now? the current batch is a long way from this afaik
Maybe. I have a fair amount of uncertainty about the speed of AI development but I think that that is well within the realm of possibility (and definitely a possibility developers should be considering). Note that even if AI replaces an entire dev team, it's still not apparent that the process would be as easy as you're saying (at least for a while after that), since after all even with a crack dev team products are almost never as easy "make me this product" "okay!" and then the product is created.
But that wasn't what I was trying to get at. My point is that this is what the author was predicting, and if that were to pass, that is more or less the death of software development as a profession, contrary to what the author says when he says "I'm not going to do that, because I absolutely don't believe it."
Indeed, but if we reach that point then we've probably got AGI, so not really much to worry about. Is there a point where AI can replace entire software dev teams, but nothing else? ie not quite AGI, seems unlikely, if you ask a product manager, they'd say Impossible! I look forward to the singularity, but I don't think this is it sadly. LLM's are a neat tool, and dev will probably change but I think it's wishful thinking on the part of business folk. the other argument is that LLM's will actually increase the software dev work, cause things that weren't possible now are which is something I find interesting.
The other thing I've been thinking is that most corporates now are mainly software (so it's been said), if thats the case and software becomes cheaper the barrier to entry to compete with corporates lowers, they become ripe for disruption. Insurance comes to mind, same with banking, probably others, search? its already disrupted, new industries will probably arise to - data validation, for example, is going to be an issue in the age of AI. The idea that making web sites for a living for the next century was probably always a very wishful way of thinking, but the idea that software development will disappear is also naive imho. However it's yet to be proved that software dev is cheaper with AI.
Isn't a big part of it knowing what to ask Claude Code - for example I wanted some code to brighten/darken my buttons on mouseover, I asked Claude to make some code to do that, it was a bit wrong, I fixed it. I integrated the code and tested it. Now I know how to do all this, because I've been doing it for years, I could have done it without Claude but it saved me a bit of time.
Now there are a few things I see that affect this
1. The only way Claude knew how to do this is because there was a stack of existing code, but its probably in C, so you could regard Claude as an expert programming language translator.
2. There is no way that Claude could integrate this into my current code base
3. Claude can't create anything new
4. It's often very wrong, and the bigger/more complex the code is, the wronger it gets.
So, what are the areas that Claude excels? it seems that CRUD web app/Web front end is the sweet spot? (not really sure about this - I don't do much web front end work). I write graphics Apps and Claude is handy for those things you'd have to look up and spend some time on, but thats about all.
An example - I asked it to make me some fancy paint brush code (to draw in a painterly style), this is hard, the code that it made was pretty bad, it just used very basic brush styles, and when pressed it went into crazy land.
So my point is - if something exists and its not too hard, Claude is great, if you want something large and complex, then Claude can be a good helper. I really don't see how it can replace a good dev, there are a lot of code monkeys around gluing web sites together that could be replaced but even then they are probably the same people who are vibe coding now.
If you really want some fun ask them to draw a circuit diagram for a simple amplifier, it's almost painful watching them struggle.
reply