OT: New Yorker Article On Coding in the Time of AI

[LionJim]

It probably has a “you clicked this website five times in the last month, you cheap bastard” paywall, but it worked for me. Very thought provoking.

A Coder Considers the Waning Days of the Craft

Coding has always felt to me like an endlessly deep and rich domain. Now I find myself wanting to write a eulogy for it.

www.newyorker.com

 

[Obliviax]

I've played with GPT 4 and attended a 3 day seminar on how it will affect my industry. I came away shocked.

We've seen these leaps in tech before: Personal computers, Macs, Eudora/Outlook, Lotus123/Excel, iPhone, app store...etc. Excel and Word put tons of medium/low skilled people out of jobs. In a few years, it settles out and becomes table stakes. They are tools, just like a shovel. Everyone has a shovel. What they do with it becomes the separator. But that is down the road. The early adopters will prosper.

Two things jumped out at me. One consultant said that AI, right now, is like that over-eager employee who never admits he/she doesn't know or needs help. So it often makes stuff up or finds data points that are in error. His point is that editing will be the key success factor. In programming, I think this is X 2. AI will make you much more efficient in getting out code, but someone has to edit it to debug and ensure the end product meets business goals. Also to make sure that it is secure. The other thing is that AI is based on history. It simply doubles down on what has always been as opposed to what will or should be. At one point, I knew a bunch of people that built a credit score based on the applicant's history. With mortgage loans, it also took into account the address. It didn't know the applicant's age, race, or anything else other than numbers (income, debt, assets). That company got sued out of existence for redlining (racist lending practices) when the whole goal was to eliminate the possibility of redlining. Turns out, it was more racist than people! Why? There was a major development funded in a bad section of town that was underwritten by the govt. But historically, it was a bad address. So the numbers came up that it was a bad place, bad investment, and not a good place to lend. My only point is that the numbers new history but did not predict the future. And AI is better, but the same. The consultants are warning about AI having racist, sexist, or partisan views based on the bulk of historical data fed into the system.

 

[johnmpsu]

One important thing to remember is that the old adage is more true today than ever:
Garbage In Garbage Out
My fear is people won't appreciate that.

The article clearly suffers from the publishers refrain, "I need 400 pages!!" but its the New Yorker so that's to be expected. There is an awful lot of drivel wedged into a few simple ideas.

I wrote software for 40 years and what separated the "good" people from the "bad" ones was talent not knowledge of the idiosyncrasies of any particular language. AI in its current state will be a helpful tool best used by talented people.

 

[Obliviax]

One important thing to remember is that the old adage is more true today than ever:
Garbage In Garbage Out
My fear is people won't appreciate that.

The article clearly suffers from the publishers refrain, "I need 400 pages!!" but its the New Yorker so that's to be expected. There is an awful lot of drivel wedged into a few simple ideas.

I wrote software for 40 years and what separated the "good" people from the "bad" ones was talent not knowledge of the idiosyncrasies of any particular language. AI in its current state will be a helpful tool best used by talented people.

Great point. While AI will make "good" faster it will also make crap faster.

 

[LionJim]

Great point. While AI will make "good" faster it will also make crap faster.

The primary point the author made is that if you have a knowledgeable, talented programmer using AI as a tool, you can do some very good work.

 

[[email protected]]

it a tool so far nothing more

 

[ManxomeLion]

Definitely interesting and thought provoking. I work in the industry and often joke about "the good news is we've done everything you've asked us to do, the bad news is we've done everything you've asked us to do".

We've had technological revolutions in the past, often with a fair amount of upheaval and change. The Industrial Revolution, Mass Media, Computers and the Internet and now AI.

While I realize there are many kinds of developers out there, there are a plethora of what I call "order takers", they're the short order cooks of coding that will turn things around regardless of whether it makes sense, is scalable or supportable going forward. AI will likely fuel this trend, and as Obliviax pointed out, there will be unintended consequences as "can we do this" and "should we do this" are often completely different conversations.

I had to laugh when the author stated near the end "A friend of mine has called this A.I. moment “the revenge of the so-so programmer.” As coding per se begins to matter less, maybe softer skills will shine." The first part I think is very true, as for the latter not so much.

I'm reminded of Edward R Murrow's "Wires and Lights in a Box" speech.
"This instrument can teach, it can illuminate; yes, and even it can inspire. But it can do so only to the extent that humans are determined to use it to those ends. Otherwise, it's nothing but wires and lights in a box."

 

[Obliviax]

Definitely interesting and thought provoking. I work in the industry and often joke about "the good news is we've done everything you've asked us to do, the bad news is we've done everything you've asked us to do".

We've had technological revolutions in the past, often with a fair amount of upheaval and change. The Industrial Revolution, Mass Media, Computers and the Internet and now AI.

While I realize there are many kinds of developers out there, there are a plethora of what I call "order takers", they're the short order cooks of coding that will turn things around regardless of whether it makes sense, is scalable or supportable going forward. AI will likely fuel this trend, and as Obliviax pointed out, there will be unintended consequences as "can we do this" and "should we do this" are often completely different conversations.

I had to laugh when the author stated near the end "A friend of mine has called this A.I. moment “the revenge of the so-so programmer.” As coding per se begins to matter less, maybe softer skills will shine." The first part I think is very true, as for the latter not so much.

I'm reminded of Edward R Murrow's "Wires and Lights in a Box" speech.
"This instrument can teach, it can illuminate; yes, and even it can inspire. But it can do so only to the extent that humans are determined to use it to those ends. Otherwise, it's nothing but wires and lights in a box."

Agreed. But it is far more about hyper-efficiency than creativity. I really compare it to Excel or Lotus123. Today, people have massive spreadsheets that a) put many people out of work and b) have made decision-making exponentially faster and more accurate. But to your point, **** in/**** out. The real winners will be those who "skate to where the puck is going to be".