| ▲ | nottorp a day ago |
| You "AI" enthusiasts always try to find a positive spin :) What if I had trusted the code? It was working after all. I'm guessing that if i asked for string manipulation code it would have done something worth posting on accidentally quadratic. |
|
| ▲ | noisy_boy a day ago | parent | next [-] |
| Depends on how toxic the culture is in your workplace. This could have been an opportunity to "work" on another JIRA task showing 600% improvement over AI generated code. |
| |
| ▲ | nottorp 17 hours ago | parent [-] | | I'll write that down for reference in case I do ever join an organization like that in the future, thanks. 600% improvement is worth what, 3 days of billable work if it lasts 5 minutes? | | |
| ▲ | noisy_boy 10 hours ago | parent [-] | | Series of such "improvements" could be fame and fortune in your team/group/vertical. In such places, the guy who toots the loudest wins the most. | | |
| ▲ | nottorp 9 hours ago | parent [-] | | So THAT's why large organizations want "AI". In such a place I should be a very loud advocate of LLMs, use them to generate 100% of my output for new tasks... ... and then "improve performance" by simply fixing all the obvious inefficiencies and brag about the 400% speedups. Hmm. Next step: instruct the "AI" to use bubblesort. |
|
|
|
|
| ▲ | FeepingCreature 17 hours ago | parent | prev | next [-] |
| > What if I had trusted the code? It was working after all. Then you would have been done five minutes earlier? I mean, this sort of reads like a parody of microoptimization. |
| |
| ▲ | nottorp 17 hours ago | parent [-] | | No, it reads like "your precious AI generates first year junior code". Like the original article. | | |
| ▲ | FeepingCreature 15 hours ago | parent [-] | | There is nothing wrong with first year junior code that does the job. | | |
| ▲ | nottorp 14 hours ago | parent [-] | | Does not. Do you know my requirements? This is actually in a time critical path. | | |
| ▲ | FeepingCreature 14 hours ago | parent [-] | | Well, that wasn't in your comment. :P If you hadn't told me that I would also not have bothered optimizing syscalls. Did you tell the AI the profiler results and ask for ways to make it faster? | | |
| ▲ | nottorp 13 hours ago | parent [-] | | > Well, that wasn't in your comment. :P Acting like a LLM now :P > Did you tell the AI the profiler results and ask for ways to make it faster? Looking for ways to turn a 10 minute job into a couple days? | | |
| ▲ | FeepingCreature 9 hours ago | parent [-] | | AI actually doesn't really work for the "a couple days" scale yet. As a heavy AI user, this sort of iterative correction would usually be priced in in a 10-minute AI session. That said- > Acting like a LLM now :P Hey, if we're going to be like that, it sure sounds like you gave the employee an incomplete spec so you could then blame it for failing. So... at least I'm not acting like a PM :P |
|
|
|
|
|
|
|
| ▲ | bongodongobob 20 hours ago | parent | prev [-] |
| Why would you blindly trust any code? Did you tell it to optimize for speed? If not, why are you surprised it didn't? |
| |
| ▲ | nottorp 17 hours ago | parent | next [-] | | So, most low level functions that enumerate the files in a directory return a structure that contains the file data from each file. Including size. You already have it in memory. Your brilliant AI calls another low level function to get the file size on the file name. (also did worse stuff but let's not go into details). Do you call reading the file size from the in memory structure that you already have a speed optimization? I call it common sense. | | |
| ▲ | miningape 15 hours ago | parent [-] | | Yep exactly, LLMs blunder over the most simple nonsense and just leaves a mess in their wake. This isn't a mistake you could make if you actually understood what the library is doing / is returning. It's so funny how these AI bros make excuse after excuse for glaring issues rather than just accept AI doesn't actually understand what it's doing (not even considering it's faster to just write good quality code on the first try). | | |
| ▲ | nottorp 14 hours ago | parent | next [-] | | The "AI" are useful for one thing. I had no idea what functions to use to scan a directory in a native C++ Windows application. Nor that they introduced an abstraction in C++ 2017?. They all work the same (needless fs access should be avoided no matter the OS) but it did give me the names*. Stuff that google search from 10 years ago would have done without pretending it's "AI". But not google search from this year. * it wasn't able to simply list the fields of the returned structure that contained a directory entry. But since it gave me the name, i was able to look it up via plain search. | | |
| ▲ | miningape 14 hours ago | parent [-] | | Yeah I find myself doing that too, use the AI to generate a bunch of names I can put into google to find a good answer. I also think if google hadn't gotten as sh*t as it has AI wouldn't be nearly as useful to most people. |
| |
| ▲ | bdhcuidbebe 7 hours ago | parent | prev [-] | | > It's so funny how these AI bros make excuse after excuse for glaring issues rather than just accept AI doesn't actually understand what it's doing Its less funny when you realize how few of these people even have experience reading and writing code. They just see code on screen, trust the machine and proclaim
victory. |
|
| |
| ▲ | johnnyanmac 17 hours ago | parent | prev [-] | | >Why would you blindly trust any code? because that is what the market is trying to sell? |
|
