I was having a discussion with my wife about the use of AI in higher education, and she suggested that students are using AI to do the homework in my finance class.
Speaking as somebody who can code, Jared, there is a lot of thought and possibly art that goes into writing robust high performance code, but if you just wants something that works and is not particularly resilient (breaks easily), then AI can certainly quickly give you that in any language.
I do agree that the advice given to those journalists was not great.
"There is no art to writing code; you just want it to work..."
OMG
Is this provincialism - being so sure that what's true in your domain is true in others?
If the code doesn't have anyone's lives depending on it, and sub-optimal operation can be repaired with a phone call or a simple legal settlement, then I might ag....no. I can't. Not even then.
I received only one A- in my undergraduate major - that was due to being six weeks late in turning in one of the three required FORTRAN 77 programs. Both the teaching assistant and I needed that long to find the error, looking one hour a week. I had typed a "B" instead of a "C" in the middle of two hundred lines of code while writing in the professor's required minimalist coding style. Yeah, never again. (Folks - use descriptive variable/function/object names and even if it means multiple physical lines are needed to make one line of code, do it. There's another rule about breaking up one's code so that parts can be effectively tested, but I better stop here.)
The month before I retired, I was asked to debug code that I had written seventeen years earlier.
(Yes, seventeen.) Because coding can be done artfully, I was able to review my code with the colleagues who asked me to debug, to review the requirements changes that had been made with them, and to have them agree in half an hour that one of the requirements changes after I left that group had been incompletely implemented.
OK, maybe that group would have given me the chess clock as a retirement present anyway...
but gosh, writing code is more than "just want[ing] it to work". There's an art to making code that works and that other people can use. Including your own self seventeen years later...
As well as not thinking, very few of our youth and young adults are serious readers. I recall as a HS senior, calculating that I had read well over a thousand books. Two books per week since second grade would have put me at a thousand, and I read significantly more than that through that phase of my life. And as a consequence of my reading, I recall MANY times where I would recognize or comprehend a concept that blew me away - that caused me to stop and ponder, and then reread what I had just read.
Those who use AI for homework, and indeed for all similar assignments at work or wherever, will never develop NEW things. Sir Isaac Newton is credited with the quote: If I have seen farther than others, it is because I stood on the shoulders of giants.
But if he had not studied and comprehended the work and ideas of those giants, he would have had no basis for developing calculus or the laws of motion, among his other contributions. Human comprehension requires work and thought, and is generally NOT something we can gain by reading an AI summary of a topic.
Perhaps what I value most by the reading and thinking I have done is a limited capacity to evaluate data and arguments (and propaganda and info/misinformation) to discern what is likely to be true and what is probably not. I fear for the future of our country/society as fewer and fewer people do the work and make the effort to likewise develop these skills.
Certainly, we should NEVER assume that an AI is telling us the truth!
Absolutely there are lazy people who will use AI to avoid the hard work of thinking, I already see it every day with some of the people I work with. At the same time, there is definitely a place for AI, especially when it comes to drudge work.
Coding is a good example. I know how to code in Python and JavaScript (among a few other languages), but it's not something I do frequently so whenever I have a project to do I find I am rusty. But ... if I know what I need to build and have an architecture in mind, I can use AI tools to dramatically speed up my rate of progress, at least to get a working prototype. Why wouldn't I spend $20 a month for something that increases my productivity? Do I need to put work into defining tests and validating the code? Sure. Do I need to review the code? Yes. It's still faster than writing the code from scratch.
To me, the key is to identify where you are adding value. You only have so many hours in a day, so try to use those as productively as you can. If AI tools help you do your job faster and better, use them. Just don't let them replace the part where you truly add value.
I'll admit to getting a little lazy/rusty in the brain dept, but not my second brain-my gut. When I see any professional, I want to know what their gut is telling them, along with objective data. Science is replete with data on this "second brain" of ours and how it works....if AI were to ever replace it, we would be doomed.
I’ll counter you somewhat here and note that AI allows me to handoff mundane work and
focus on what requires greater thought. I can spend more time on ideation and less time a writing requirements. I can craft a consistent prompt and append my ideation to produce a well crafted set of requirements. The results may require tweaks but the AI results get better with each cycle. What was 1/2 day of effort has been reduced to 1/2 hour leaving me more time to think deeply about the next issue.
Speaking as somebody who can code, Jared, there is a lot of thought and possibly art that goes into writing robust high performance code, but if you just wants something that works and is not particularly resilient (breaks easily), then AI can certainly quickly give you that in any language.
I do agree that the advice given to those journalists was not great.
“Thinking is the hardest work there is, which is the probable reason why so few people engage in it.” Henry Ford.
Great piece
BTW Jelly Roll doesn’t take the pill. All hard work. He says it would affect his voice.
"There is no art to writing code; you just want it to work..."
OMG
Is this provincialism - being so sure that what's true in your domain is true in others?
If the code doesn't have anyone's lives depending on it, and sub-optimal operation can be repaired with a phone call or a simple legal settlement, then I might ag....no. I can't. Not even then.
I received only one A- in my undergraduate major - that was due to being six weeks late in turning in one of the three required FORTRAN 77 programs. Both the teaching assistant and I needed that long to find the error, looking one hour a week. I had typed a "B" instead of a "C" in the middle of two hundred lines of code while writing in the professor's required minimalist coding style. Yeah, never again. (Folks - use descriptive variable/function/object names and even if it means multiple physical lines are needed to make one line of code, do it. There's another rule about breaking up one's code so that parts can be effectively tested, but I better stop here.)
The month before I retired, I was asked to debug code that I had written seventeen years earlier.
(Yes, seventeen.) Because coding can be done artfully, I was able to review my code with the colleagues who asked me to debug, to review the requirements changes that had been made with them, and to have them agree in half an hour that one of the requirements changes after I left that group had been incompletely implemented.
OK, maybe that group would have given me the chess clock as a retirement present anyway...
but gosh, writing code is more than "just want[ing] it to work". There's an art to making code that works and that other people can use. Including your own self seventeen years later...
As well as not thinking, very few of our youth and young adults are serious readers. I recall as a HS senior, calculating that I had read well over a thousand books. Two books per week since second grade would have put me at a thousand, and I read significantly more than that through that phase of my life. And as a consequence of my reading, I recall MANY times where I would recognize or comprehend a concept that blew me away - that caused me to stop and ponder, and then reread what I had just read.
Those who use AI for homework, and indeed for all similar assignments at work or wherever, will never develop NEW things. Sir Isaac Newton is credited with the quote: If I have seen farther than others, it is because I stood on the shoulders of giants.
But if he had not studied and comprehended the work and ideas of those giants, he would have had no basis for developing calculus or the laws of motion, among his other contributions. Human comprehension requires work and thought, and is generally NOT something we can gain by reading an AI summary of a topic.
Perhaps what I value most by the reading and thinking I have done is a limited capacity to evaluate data and arguments (and propaganda and info/misinformation) to discern what is likely to be true and what is probably not. I fear for the future of our country/society as fewer and fewer people do the work and make the effort to likewise develop these skills.
Certainly, we should NEVER assume that an AI is telling us the truth!
Absolutely there are lazy people who will use AI to avoid the hard work of thinking, I already see it every day with some of the people I work with. At the same time, there is definitely a place for AI, especially when it comes to drudge work.
Coding is a good example. I know how to code in Python and JavaScript (among a few other languages), but it's not something I do frequently so whenever I have a project to do I find I am rusty. But ... if I know what I need to build and have an architecture in mind, I can use AI tools to dramatically speed up my rate of progress, at least to get a working prototype. Why wouldn't I spend $20 a month for something that increases my productivity? Do I need to put work into defining tests and validating the code? Sure. Do I need to review the code? Yes. It's still faster than writing the code from scratch.
To me, the key is to identify where you are adding value. You only have so many hours in a day, so try to use those as productively as you can. If AI tools help you do your job faster and better, use them. Just don't let them replace the part where you truly add value.
I'll admit to getting a little lazy/rusty in the brain dept, but not my second brain-my gut. When I see any professional, I want to know what their gut is telling them, along with objective data. Science is replete with data on this "second brain" of ours and how it works....if AI were to ever replace it, we would be doomed.
Did you have Ai do the research on Jelly Roll? He did the hard work himself without the poll as the other person commented.
Simple research would have shown that. Ai would most likely have given you the right answer.
I agree that using your brain is important. But using your hands, feet, muscles, is just as important. It’s all about balance.
I’ll counter you somewhat here and note that AI allows me to handoff mundane work and
focus on what requires greater thought. I can spend more time on ideation and less time a writing requirements. I can craft a consistent prompt and append my ideation to produce a well crafted set of requirements. The results may require tweaks but the AI results get better with each cycle. What was 1/2 day of effort has been reduced to 1/2 hour leaving me more time to think deeply about the next issue.
Great article.
Not using AI would be like not using a calculator.
Those who learn to wield AI won’t just get the job done, they’ll leverage it to 10x improved outcomes.