Next you'll tell me I shouldn't use a calculator, or a computer at all.
The conceptual level at which I work benefits massively from the recent developments in LLMs, and to stop training myself for the new meta means to drastically fall behind and possibly miss my goals. There is absolutely no reason not to evolve alongside this new technology.
Imagine that the calculator breaks, and you need to do some quick math. (Yeah, yeah, I know. Cell Phones and such.) It is good to know how to do math without a calculator for those edge cases where one is not available, or when you can't do the equations that you need with it.
Calculators are good. Calculators are useful. Calculators accelerate your workflow beyond what your ancestors could do. Not knowing how to do math without one is still a hindrance, hence why we still need math classes. You need to know the underlying theory of why the calculators do what they do in order for them to be useful.
It's the same with ChatGPT. ChatGPT is a fantastic tool that can benefit your workflow. I use it all the time myself. However, being 100% dependent on it for work is a dangerous game. If it goes down(like today) or the company behind it makes a change that makes it less useful, you still need to know how to do your job without it. That's why OP's comment is worrying. They said that they feel unable to work without it. It reminds me of that Avengers quote: "If you are nothing without the Iron Man suit, you shouldn't have it."
The point of my previous comment was that the kind of work I do is not "quick math" and taking GPT out of the equations reduces my velocity by magnitudes. Calculators aren't going to disappear and no astrophysicist is going to do massive multi-dimensional calculations by hand.
> If you are nothing without the Iron Man suit, you shouldn't have it.
It's a nice thought, but I can apply this chain of reasoning to no end of technologies without which scientific progress in a given domain would entirely halt.
> Taking GPT out of the equations reduces my velocity by magnitudes.
But you should still know how to do it regardless. Because of situations like this. That is the point of my comment. If you can't do your job without ChatGPT, you don't have any business working in your field to begin with. Even if it's at a reduced speed, you still need to know how to do your job.
>It's a nice thought, but I can apply this chain of reasoning to no end of technologies without which scientific progress in a given domain would entirely halt.
Not really. To do advanced stuff you have to understand the basics. This goes for almost every field. You can't build the next-level Javascript app without knowing what an if-else does. You can't be a doctor without knowing a little chemistry and biology. Even in a job like construction, you need to be able to do simple math to make sure your measurements are correct.
Saying that advanced tools should be used for things like programming without understanding the basics is a logical fallacy. It's the same argument that managers sometimes use. You know, the "programmers only copy and paste from stack overflow. Why do we pay you so much?" Asking chatGPT for code means nothing if you don't know how to apply it and search for bugs. And to use code from ChatGPT, you need to know how to do your job without it. Otherwise, you will only produce code that, at best, sucks and, at worst, doesn't work.
> But you should still know how to do it regardless. Because of situations like this. That is the point of my comment. If you can't do your job without ChatGPT, you don't have any business working in your field to begin with. Even if it's at a reduced speed, you still need to know how to do your job.
I'm currently using ChatGPT for a bunch of AI/ML that I don't know how the insides are working. But I'm able to build models from scratch that does exactly what I want, with 99% accuracy in my test cases, without actually knowing what the model does, but together with GPT4 + automatic hyperparameter tuning, I'm able to build models I can use in production.
Does it matter if I know exactly how everything inside in the model works, if I can get it to work exactly to my specification without it?
This is essentially how I started programming as well way back in time. I didn't know exactly what the Perl code I copy-pasted did, but if it solved the problem, it solved the problem. It brought me and my family out of poverty, and at that point I couldn't care less about how the magic actually was done, just that it did work.
Obviously now I have more knowledge about web field in general and 10+ languages that I no longer have to use any docs to be productive with, and maybe that'll happen with AI/ML eventually as well, but for a person who is starting with something new and wanna be productive quickly, GPT4 is a godsend.
>Does it matter if I know exactly how everything inside in the model works, if I can get it to work exactly to my specification without it?
Maybe not every single thing, but you should know what your lines of code that implement it do. You can't debug if you don't know why you wrote what you wrote.
>This is essentially how I started programming as well way back in time. I didn't know exactly what the Perl code I copy-pasted did, but if it solved the problem, it solved the problem. It brought me and my family out of poverty, and at that point, I couldn't care less about how the magic actually was done, just that it did work.
Ok. Great that it pulled you out of poverty. That's irrelevant to your argument, but Im glad for you. I guarantee you that the code sucked regardless. You might not care that you produced software that sucked, but it still guaranteed sucked. If you copy and paste code without knowing what it does, you are a bad programmer. Someone, somewhere, is going to have to clean up your mess. And they are cursing your name right now.
> Not really. To do advanced stuff you have to understand the basics.
Doesn't that defeat the whole purpose of these abstractions? Can you read the machine code that your C compiler produces? How much of electrical engineering do you need to know to write a bash script? The physics of how a NAND gate is implemented?
It's obviously in the early stages, and I don't disagree with you completely today -- but this will just be one more layer on top of an already deep stack of abstractions that underlie all of computing.
>Can you read the machine code that your C compiler produces? How much of electrical engineering do you need to know to write a bash script? The physics of how a NAND gate is implemented?
Unironically, I can and do. I took classes in college for things like ASM and logic gates. That's not the point I was making, though. The point is that you need to know how to read your code so that you can fix it if you have to maintain it. Or, if ChatGPT is down(like today) or not giving you the right answer, you can still do your work, albeit a little more slowly. My worry is that people will just plop whatever into a compiler, and leave buggy code that introduces bugs and security vulnerabilities. An LLM is only as good as its data source, and with things like ChatGPT and Github copilot, that data source is programmers both experienced and inexperienced. Use it, love it, but don't rely on it. Implement best practices, and use your head.
The calculator is not a service but a tool. Until LLMS don't become just a tool don't rely on them too much or expect it be broken and have a workaround for that.
I have backup local LLMs which I used during the outage. This doesn't prevent the fact that for now, ChatGPT-4 wins out in output quality.
This won't remain true for long and so it is actually harmful to my career to not invest time learning how to use these tools now, instead of waiting for the time when they are perfect.
ChatGPT is the black box that is pushing the buttons of the calculator for you. You don’t learn maths or programming with this service.
If the calculator is broken, I can still work slowly but I understand what I do. Without experience, you can’t understand what the black box is giving you.
This is why wherever I travel in the world I print out paper maps of every city and village I think I might go, I don't want to get a dependency on a digital map. My paper maps have a 100% uptime.
The conceptual level at which I work benefits massively from the recent developments in LLMs, and to stop training myself for the new meta means to drastically fall behind and possibly miss my goals. There is absolutely no reason not to evolve alongside this new technology.