It’s fascinating how many of you use LLM’s but “of course check the results”.
Doesn’t this make it more work than if you’d just, you know, do it yourself?
It’s fascinating how many of you use LLM’s but “of course check the results”.
Doesn’t this make it more work than if you’d just, you know, do it yourself?
@thomasfuchs I've had a number of coworkers use them to generate templates like risk models, SteerCo charters and things like that.
I've used it to brainstorm project names because I've found catchier names get more funding and attention
@SirDarkStar No, it’s not similar.
Unit tests compare input and output values, code transforms input into output values.
These are two diametrically opposite operations.
@thomasfuchs isn’t that similar to asking “why write unit tests if you are writing the code”? We also do code reviews because a lot of people are worse than LLMs :)
As others mentioned it basically comes down to being worth it when the cost of synthesis is higher than the cost of validation/fixing.
Also, we’re mostly all in learning mode I think.
@Fryguy Yes, that's true for most learning.
People learn by doing, not by seeing.
@thomasfuchs I had a code LLM turned on in vscode while I was learning a new language from a book and it was autofilling in the code for me. At first I thought it was cute how it knew the book contents, but then I realized it was actually robbing me of the actual learning process. Turns out the act of writing out those examples myself was important for the way I learn.
GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.
All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.