Coding Jobs and GPT-4 | Hacker Information

A few thoughts based on my experience with GPT4 so far.

1. For now, not by much. It will make us more productive but the world is not short of projects with endless backlogs so huge nobody can ever drain them. Maybe some of those tickets are low value or obsolete, but a lot of them represent real bugs or ideas for improvement that the cost of software development makes prohibitive to reach.

2. It may change the nature of the job a lot. It could hurt developers who are very heads down and ideas constrained. If what you love most about coding is the thrill of finding the perfect algorithmic implementation, then, well, AI may reduce the enjoyment of the job. If what you love about coding is seeing your ideas come alive, then AI can increase your enjoyment of the job.

3. Whilst some people claim they ideate and get inspired by talking to AI, my own experience has been that GPT4 is a rather conservative and predictable sort of personality. Perhaps it’s the training and maybe via the API you can ramp up the temperature and get more creativity out of it, but whenever I’ve asked it for ideas or tried to bounce ideas around, I tend to get milquetoast yes-man type results, or the ideas it comes up with are generic and obvious. Also with each day that passes it’s getting easier to notice the lack of “AI invents/solves something new” type stories. There was the one where it came up with a word game, which was pretty cool, and we know it can invent stories. But it doesn’t yet seem to be producing an abundance of new ideas for things like new features, business innovations, etc. Maybe it does for other people. Maybe it will start doing it for me soon. But for now, it doesn’t seem able to do that.

So – programmers who just want to play code golf, be ready for that to become more of a hobby than a job. But for programmers who always wished they had more hours in the day to get through all their ideas, it will enable them to create a lot more value and that value will in turn create demand for yet more value-add on top of that. So it should be a virtuous circle, in theory.

Yes it absolutely decimate demand for junior coders.
My firm has basically frozen hiring, and all existing devs are expected to boost their productivity using one of the all the various AI’s or get fired.

We expect the economy to get way worse as de-dollarization seems to be looming on the horizon, and preparing now.

while chatgpt and bing are useful, they absolutely do not replace the pipeline of jr devs one needs to someday have sr devs with deep domain knowledge. IMHO it’s a great ‘search’ technology, but it shouldn’t be taken for granted or assumed correct.

Also, re de-dollarization. This pops up every 5 years or so when some autocratic regime gets tired of operating under the us dollar hegemony, but I don’t see any other options that don’t have their own major issues. The euro? That currency is fraying at the seams from trying to stuff high and low performing countries (i.e. Germany and Italy) together without enough control over each country’s finances. China? They’ve got to free float and stop manipulating the yuan first.

Literally nobody is as trustworthy as the USD, no matter how badly Moscow, Beijing or Tehran might wish it to not be so.

I was afraid to hear stories like this. It’s already pretty bad that less and less companies are willing to take the (originally completely obligatory) burden of teaching new junior programmers. It feels like we really are doing way too little to raise any next generation of any shape.

As a ‘professional coder’, I’ve been very much pondering – and experimentinng – with how best to use AI coding tools.

For big, complex, programs I work on (multiple C programs up to 100k LoC) it’s hard to frame a question in such a way that it doesn’t need to know details of the codebase. For peripheral stuff it’s pretty useful as a quick way to get some cut’n’paste code (‘write me code to read the load average on linux and log it to a file every second’ – I know exactly how to do that, but it’s a faster typer).

So initial take is it’s a very useful new tool – so sort-of a corollary, people who learn how to use it quickly and well are going to have an advantage in the short term. And even if/when everyone uses it, some people are going to be better at using it (i.e. asking the right question – like I find myself groping around for, when asking it about UK company law earlier today).

I don’t know what it will look like, but remember that the job of a “coder” is not to “write code” but to solve business problems.

The programming profession has always demanded that developers operate at higher and higher levels of abstraction and produce more/bigger/faster. I expect LLMs to continue that trend – a tool to get more done at a bigger scale.

Personally, I think no-code tools are much more of an existential threat than LLMs. Or rather, they should be. Instead, despite the prevalence of tools that seek to democratize software authorship, the demand for software engineers has yet to abate.

As for why, my best theory is that writing code isn’t the hard part, but in fact one of the easier parts of the job, much like how drafting isn’t really the hard part when it comes to engineering or architecture.

Writing code is one thing, owning it and having responsibility is another. But can we build tooling for this to on top of LLMs too? This won’t happen overnight but it will certainly happen gradually, I don’t see why not. By owning it, I mean, upgrading libraries, fixing bugs, writing tests, running tests, fixing tests, and maybe even adding features.

It’s totally possible to make LLMs write actually good code that can be maintained using normal software engineering practices.

Of course there will still be a need for a human touch, such as knowing limitations of LLMs and possible workarounds and understanding business requirements. Such productivity boost may even create more jobs as products will become cheaper and faster to create, hence more new markets will be explored. The pie might also get bigger not smaller.

The question I ask myself: Is there a large and still-unmet demand for software engineering worldwide. AKA is the market supply-constrained by a large-enough amount?

If so, I imagine engineers will be more productive with AI (and wages will go down, or at least stop growing somewhat), but the demand for software engineers will stay strong.

If not, then engineers being more productive would mean fewer engineers can meet the global demand for software engineering work and I would expect to see the demand for software engineers reduced.

Development has gotten much faster since the 2000s: better languages, better IDEs, better libraries, better design practices, better documentation, better online resources, better (faster, less buggy) target platforms. I remember when I was first writing code in the early 2010s writing Objective-C in Xcode 3 with manual memory management. Developer productivity has exploded more than most people realize.

And yet developers have always been in demand. We just have a lot more programs now, and a lot of very-similar and/or niche programs. All the random libraries on npm, Rust, Haskell, etc. the various crypto-currencies, static/dynamic webpage builders, 3 separate JavaScript runtimes. People start businesses and actually get funding for these libraries. And it seems like a lot of companies want almost the exact same product, some “business solution” or “cloud solution”, but they have specific reasons existing solutions aren’t acceptable (performance? security? some feature?), so they pay $200k+ salaries for developers to build them.

But will this always be the case? Even ignoring GPT4, we don’t really need these developers who are working on niche and similar projects; we are having a harder time getting hired right now with the economic downturn. And still, it’s entirely possible this growth has a limit, and GPT4 will make developers turn ideas into working products faster than people can come up with them.

Actually I’ve found the “React monkeys” you dismiss to be the people who work with product & UX on what the actually product being developed should do. The “backend people” who write API plumbing and DB tables are the ones I’d worry about.

i might be wrong but i disagree with a lot of commenter’s saying how they are not hiring juniors anymore or their seniors are 10x more productive even if they are just boilerplate programmers they are still more useful than tools like gpt and always will be (my opinion) my guess is that instead of juniors they will be called coding assistants.

Are there fewer engineers since AutoCAD was invented and they didn’t need to draw everything out by hand? Yes, but not by much and they just have much higher expectations.

I feel like this is more programmers realizing they have to become the 10x they were told was a myth to survive. Which is kinda true! BUT what also is true using AI tools will make them that if they learn to use them, now they need to know security inside and out, systems design etc. Higher level concepts way above a code monkey’s pay grade, now will be expectations, all the theory side is going to be much more important.

It’s a strange question. That “professional coder” term is such a misconception… i mean does anyone know of any experienced dev that consider themselves a coder?

Large language models will change the way we write code, but not make code disappear. We will always need code to execute software.

But we don’t use punched cards anymore. We don’t write low level machine code anymore, unless we must or we are into that. We don’t have to use complex and difficult programming languages when we can do the same in a few lines of Python that abstract everything. And now we don’t have to write most of the high level code as we can have a loose conversation with the machine.

I’m expecting that the demand for good coders will stay high, to use these new AIs, fix the bugs, check the outputs, and build much more ambitious projects.

However the demand for cheap coders that write simple applications will disappear and be replaced by machines. The humans will do other activities instead, in software development or not.

What's Your Reaction?

Leave a Reply