Conversation

Notices

1. Embed this notice
   Jason Gorman (jasongorman@mastodon.cloud)'s status on Friday, 26-May-2023 18:05:47 JST Jason Gorman

   We spend roughly 10x as much time reading code as we do writing it. A tool or technique that makes you twice as "productive" at writing code *at best* makes you 5% more productive over all. Making your code easier to understand will have 10x the impact. But that doesn't sell tools or put developers out of work, so you won't be reading about it in Forbes.

   • clacke likes this.
   • clacke repeated this.
   • Embed this notice
     Kees de Kooter 🍉 (kdekooter@mastodon.social)'s status on Sunday, 28-May-2023 14:20:54 JST Kees de Kooter 🍉

     in reply to

     @jasongorman and all this AI-generated code needs to be reviewed. There goes your 5% time saving...

     clacke likes this.
   • Embed this notice
     Frank Quednau (flq@freiburg.social)'s status on Sunday, 28-May-2023 14:22:18 JST Frank Quednau

     in reply to

     @jasongorman that it doesn’t sell tools I’m not so sure. A tool that would help me ensure readable code, however that would work, I would find a very interesting proposition

     clacke likes this.
   • Embed this notice
     Jason Gorman (jasongorman@mastodon.cloud)'s status on Sunday, 28-May-2023 14:22:20 JST Jason Gorman

     in reply to

     • Frank Quednau

     @flq We might be waiting a while for that 🙂

     clacke likes this.
   • Embed this notice
     Jason Gorman (jasongorman@mastodon.cloud)'s status on Sunday, 28-May-2023 14:22:22 JST Jason Gorman

     in reply to

     "But Jason, we can get ChatGPT to explain the code to us". Oh really?

     clacke likes this.
   • Embed this notice
     Chris Pitts (thirstybear@agilodon.social)'s status on Sunday, 28-May-2023 14:25:45 JST Chris Pitts

     in reply to

     @jasongorman Obviously ChatGPT is a master of stating the bleedin’ obvious.

     clacke likes this.
   • Embed this notice
     Chris Pitts (thirstybear@agilodon.social)'s status on Sunday, 28-May-2023 14:25:48 JST Chris Pitts

     in reply to

     @jasongorman Yes, but I have worked with (senior) developers who commented their code just like that 😁

     clacke likes this.
   • Embed this notice
     Jason Gorman (jasongorman@mastodon.cloud)'s status on Sunday, 28-May-2023 14:25:49 JST Jason Gorman

     in reply to

     • Chris Pitts

     @thirstybear That's the thing: if it doesn't have the information, it can't generate an explanation. It excels at explaining code that's easy to understand. The real selling point is brownie points for comments. Like at university.

   • Embed this notice
     Jason Gorman (jasongorman@mastodon.cloud)'s status on Sunday, 28-May-2023 14:25:50 JST Jason Gorman

     in reply to

     • Julian Elve

     @julian It's actually one of the refuctorings in my Waterfall 2006 presentation: "Stating The Bleeding Obvious"

     https://www.waterfall2006.com/gorman.html

     clacke likes this.
   • Embed this notice
     Julian Elve (julian@social.synesthesia.co.uk)'s status on Sunday, 28-May-2023 14:25:58 JST Julian Elve

     in reply to

     @jasongorman classy comments there 😉

     clacke repeated this.
   • Embed this notice
     M. Forester (mforester@rollenspiel.social)'s status on Sunday, 28-May-2023 14:26:25 JST M. Forester

     in reply to

     @jasongorman I'll give it that, but it really depends on what you define as *explaining* the code.
     I saw that technically all observations were correct, but none of them are helpful. If I'd ask real intelligence (a.k.a. a person), they'd (hopefully) explain the function in more broad strokes and go deeper and deeper until I understand what I need. 🙂

     clacke likes this.
   • Embed this notice
     Jason Gorman (jasongorman@mastodon.cloud)'s status on Sunday, 28-May-2023 14:26:26 JST Jason Gorman

     in reply to

     • M. Forester

     @mforester TBF, if the information is there, it can explain the code. Basically, it can explain code that's easy to understand.

   • Embed this notice
     M. Forester (mforester@rollenspiel.social)'s status on Sunday, 28-May-2023 14:26:27 JST M. Forester

     in reply to

     @jasongorman this is a perfect example of how not to write comments. Brilliant. 🤩
     ChatGPT explains exactly zilch...as expected.

   • Embed this notice
     Lars Marowsky-Brée 😷 (larsmb@mastodon.online)'s status on Sunday, 28-May-2023 14:27:54 JST Lars Marowsky-Brée 😷

     in reply to

     @jasongorman I hate to say it (because I very much dislike the LLM/AI hype), but ... they're surprisingly good at answering questions about code one shows them too. Not perfect, but they can provide inspiration and places to start looking.
     (If one knows how to ask a question, obvs.)

   • Embed this notice
     Jason Gorman (jasongorman@mastodon.cloud)'s status on Sunday, 28-May-2023 14:27:54 JST Jason Gorman

     in reply to

     • Lars Marowsky-Brée 😷

     @larsmb I'm not so sure about that ;-)

     clacke likes this.
   • Embed this notice
     aadmaa (aadmaa@mathstodon.xyz)'s status on Sunday, 28-May-2023 14:28:24 JST aadmaa

     in reply to

     @jasongorman Your old waterfall post is hilarious; I had not seen that. For one's own sanity, a refuctoring transpiler could be handy (a sort of "enfuctor" or "fuctorizer" or "infucorator" or "undefuctorizer") so one could maintain two copies of code with minimal work, like a fraudulent business might maintain two sets of books. This is like the old code obscuring tools, except of course it should appear that the code is not *intentionally* obscure.

     Additionally, the transpiler could be designed to make outputted tests more fragile and interdependent. So all the tests still pass, but minimal changes that really should not break the tests, break substantially all of them.

     clacke likes this.
   • Embed this notice
     Kevin Trainor (_kevintrainor@mastodon.online)'s status on Sunday, 28-May-2023 14:29:23 JST Kevin Trainor

     in reply to

     @jasongorman It just occurred to me when a programmer takes the clear ideas in their head and creates code with bad variable names and function names that a lossy transformation is taking place. It is not possible for a person or an LLM to subsequently offset that loss by processing the code. The information has been lost.

     clacke likes this.
   • Embed this notice
     cryo2go :unverified: (benjaminhimes@struct.bio)'s status on Sunday, 28-May-2023 14:29:52 JST cryo2go :unverified:

     in reply to

     @jasongorman i personally spend 95% of my time shouting and shaking my fist at code

     clacke likes this.
   • Embed this notice
     Thomas Depierre (di4na@hachyderm.io)'s status on Sunday, 28-May-2023 14:31:21 JST Thomas Depierre

     in reply to

     @jasongorman I may have been writing about this stuff a bit lately... https://hachyderm.io/@Di4na/110431524694102577

     clacke likes this.
   • Embed this notice
     Jason Gorman (jasongorman@mastodon.cloud)'s status on Sunday, 28-May-2023 14:32:06 JST Jason Gorman

     in reply to

     • mnl mnl mnl mnl mnl

     @mnl They certainly generate a lot of code that *should* be thrown away ;-)

     clacke likes this.
   • Embed this notice
     mnl mnl mnl mnl mnl (mnl@hachyderm.io)'s status on Sunday, 28-May-2023 14:32:07 JST mnl mnl mnl mnl mnl

     in reply to

     @jasongorman This might be confounding multiple things. We might spend so much time reading code because: the code is hard to read and because writing code is hard too.

     I think that LLMs actually help for both cases, but they need us to rethink how we write and read code. LLMs help a lot to make code consistent stylistically (API patterns and documentation phrasing) and they can generate a lot of code that can be thrown away on the path to consistent abstractions.

   • Embed this notice
     Jason Gorman (jasongorman@mastodon.cloud)'s status on Sunday, 28-May-2023 14:32:26 JST Jason Gorman

     in reply to

     • mnl mnl mnl mnl mnl

     @mnl The only drawback I see is what I've experienced recently of junior devs interacting with those tools, unaware that it's feeding them BS. It's the same problem as copy/paste programming. *Exactly* the same, in fact.

     clacke likes this.
   • Embed this notice
     mnl mnl mnl mnl mnl (mnl@hachyderm.io)'s status on Sunday, 28-May-2023 14:32:27 JST mnl mnl mnl mnl mnl

     in reply to

     @jasongorman That approach centers the developer, instead of expecting the machine to write and comprehend code. The LLM is there to allow the developer to be more mindful about the creation and consumption of their code, as a linter for documentation and specification as well as a quick prototype playground.

     I hope we see more UX going in that direction instead of pretending that a chatbot can write something decent based on a single line prompt.

   • Embed this notice
     Jason Gorman (jasongorman@mastodon.cloud)'s status on Sunday, 28-May-2023 14:35:09 JST Jason Gorman

     in reply to

     • deSilva
     • Guus der Kinderen

     @aesthetikx @guusdk Record yourself "coding" for a day on a production system

     clacke likes this.
   • Embed this notice
     deSilva (aesthetikx@ruby.social)'s status on Sunday, 28-May-2023 14:35:10 JST deSilva

     in reply to

     • Guus der Kinderen

     @guusdk @jasongorman I was going to ask the same thing. I'm really not sure this is true, it certainly doesn't "feel" true, unless you count reading the code you are actively writing. Certainly this depends on the kind of project.

   • Embed this notice
     Guus der Kinderen (guusdk@toot.igniterealtime.org)'s status on Sunday, 28-May-2023 14:35:17 JST Guus der Kinderen

     in reply to

     @jasongorman "We spend roughly 10x as much time reading code as we do writing it."

     Interesting! Is that based on empirical data? I had not looked at it this way.

   • Embed this notice
     Graham Lee (leeg@fosstodon.org)'s status on Sunday, 28-May-2023 14:36:41 JST Graham Lee

     in reply to

     @jasongorman there's some maths missing here. If I double my writing rate then I double the amount I have to read, so I become 90% worse overall not 5% better.

     clacke likes this.
   • Embed this notice
     Jason Gorman (jasongorman@mastodon.cloud)'s status on Sunday, 28-May-2023 14:36:42 JST Jason Gorman

     in reply to

     • Graham Lee

     @leeg Bingo!

   • Embed this notice
     clacke (clacke@libranet.de)'s status on Sunday, 28-May-2023 14:36:42 JST clacke

     in reply to

     • Graham Lee
     @jasongorman @leeg Yeah. Our job is not to write code. Code in itself is not an asset. Code is a liability we take on in order to solve business problems.