When I was a kid and got an assignement for writing an essey about "why good forces prevailed in Lords of the Rings" as a gate check to see if I actually read the novel I had three choices:
(a) read the novel and write the essey myself
(b) find an already written essey - not an easy task in pre-internet era but we had books with esseys on most common topics you could "copy-paste" - and risk that the professor is familiar with the source or someone else used the same source
(c) ask class mate to give me their essey as a template and rephrase it as my own
A and C would let me learn about the novel and let me polish my writing skills.
Today I can ask ChatGPT to write me a 4 pages essay about a novel I've never heard of and call it a day. There's no value gained in the process.
That's a simple example. The problem is that the same applies to programming. Novice programmer will claim that LLM give them power to take on hard tasks and programm in languages they were not familiar before. But they are not gaining any skill nor knowledege from that experience.
If I ask google maps to plot me a directions from Prague to Brussels it will yield a list of turns that will guide me to my destinations, but by any means I can't claim I've learned topography of Germany in the process.
> Today I can ask ChatGPT to write me a 4 pages essay about a novel I've never heard of and call it a day. There's no value gained in the process.
If we take the original article at face value, no you can't do that. ChatGPT will apparently produce something that is obviously ChatGPT produced and fail to fool even the most absent minded of instructors that you have read the material. So even with a ChatGPT LLM to help you out, you're largely going to have to do a modified version of C, replacing your class mate with the LLM and adding in the need to do your own reading and validation to ensure that the text matches the actual book contents.
> If I ask google maps to plot me a directions from Prague to Brussels it will yield a list of turns that will guide me to my destinations, but by any means I can't claim I've learned topography of Germany in the process.
I would argue that even if you plotted a route by hand reading maps, you can't claim to have learned the topography of Germany either. "The map isn't the territory" after all.
ChatGPT and the like are in a weird position at the moment. It's usually pretty clear that you have written it using an LLM, but it's hard to PROVE. And you need to be able to prove it (to some degree, which varies depending on the institution) to reliably count off for it, otherwise the student will challenge your finding that they cheated and the Honor Court (or administrator, or other equivalent) will tell you that you can't do that.
So, you can usually get away with it if there is not some way the professor/TA can prove it.
As things change, this will change, but that's the situation the author of the original article finds themself in, because it's the current situation.
> If we take the original article at face value, no you can't do that. ChatGPT will apparently produce something that is obviously ChatGPT produced and fail to fool even the most absent minded of instructors that you have read the material.
Only if you don't have any custom instructions about style and don't proofread it afterwards. All the usual "tells" of ChatGPT are very obvious to scrub out, and you don't have to use OpenAI's chat wrapper to begin with.
Ok, but if you're going to proofread ChatGPT's output and edit it and massage it until you get something that isn't obviously the output of the LLM, how is that different from taking your buddy's homework and changing enough of the text to make it look like you didn't copy it? In either case you have to read what someone else wrote and comprehend it enough to decide what should or shouldn't change.
> Novice programmer will claim that LLM give them power to take on hard tasks and programm in languages they were not familiar before. But they are not gaining any skill nor knowledege from that experience.
Not true.
Using LLM to learn quickly a new programming language + being productive is best method ever. If you pay attention, you acquire rapidly new skill and knowledge, and those that are relevant to your job.
Using LLM is MUCH more efficient than reading a book going through all the minute details of the language prior telling how to use it. It's the same as learning a language from your parents compared to learning a language from a class. You might not know all the grammar rules, but you'll be way more proficient. And nothing prevents you from learning the grammar later on.
> Using LLM to learn quickly a new programming language + being productive is best method ever. If you pay attention, you acquire rapidly new skill and knowledge, and those that are relevant to your job.
I tend to disgree. maybe for you, but my mind does not work that way.
As a teacher at a university I come to see that students "learn" by asking an LLM, but they forget to understand the content the LLM produces because the LLM actually solves the assignment (mostly) for them. One may say that's the teachers task to produce better questions, but the thing I most "struggle" with that getting educated as a student seems to just be a play of "gaming the system". Yes, it was similar during my time (learning how to reach your goal with as less effort as possible is a good part of the "education" at a university IMHO), but we actually had to think and understand while today just seems like prompt-and-copy.
> If I ask google maps to plot me a directions from Prague to Brussels it will yield a list of turns that will guide me to my destinations, but by any means I can't claim I've learned topography of Germany in the process.
There are multiple ways you can use such technology, too. If you use Google Maps with its out-of-the-box configuration for turn-by-turn directions, with it oriented in the direction of travel, you won’t learn so much; but if you change it to always display the map north-up, and look at the map it shows you—inferior though it be to good paper maps, in most cases—it’s easier to develop a feel for layouts and geography.
When I was a kid and got an assignement for writing an essey about "why good forces prevailed in Lords of the Rings" as a gate check to see if I actually read the novel I had three choices: (a) read the novel and write the essey myself (b) find an already written essey - not an easy task in pre-internet era but we had books with esseys on most common topics you could "copy-paste" - and risk that the professor is familiar with the source or someone else used the same source (c) ask class mate to give me their essey as a template and rephrase it as my own
A and C would let me learn about the novel and let me polish my writing skills.
Today I can ask ChatGPT to write me a 4 pages essay about a novel I've never heard of and call it a day. There's no value gained in the process.
That's a simple example. The problem is that the same applies to programming. Novice programmer will claim that LLM give them power to take on hard tasks and programm in languages they were not familiar before. But they are not gaining any skill nor knowledege from that experience.
If I ask google maps to plot me a directions from Prague to Brussels it will yield a list of turns that will guide me to my destinations, but by any means I can't claim I've learned topography of Germany in the process.