In the garden, and RAIN!!! (2)
POSTED BY: SIGNYM
UPDATED: Monday, September 7, 2026 11:50
VIEWED: 338648
PAGE 834 of 834
Nothing like a little dancing too day to Doug and the Slugs. 
Good music.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Quote:
Originally posted by Brenda:
Nothing like a little dancing too day to Doug and the Slugs.
Doug and the Slugs???
Is that because it's Canadian and they didn't break through that I never heard of them before, Brenda? I can't believe I've ever heard that name before.
You dance at home? Good for you. I always wanted to do that, but it's really rare that I do. Sometimes if I've got enough coffee in me, enough endorphins from working outside, and maybe a project is going really well, I've been known to dance around a bit too.
But only ever when I'm inside and nobody can see me.
--------------------------------------------------
Those who dance always seem crazy to those who can't hear the music.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Man...
It's always a game of just learning how it behaves, and working your way around it.
We've gotten so much fixed, some of which we never got working before, but I'm still trying to get our main execution going after half a day of it not completing.
But now I'm finally starting to understand the why...
Just how bloated and bullshit our code was because how dumb the both of us were in the early sessions when just blindly putting it all together with no guardrails. And all those code changes that didn't fix something but were never properly backed out.
And realizing just how true that is when learning the limits of its memory... not its physical ram, but its "conversational memory" that always prioritizes the last turn over literally everything else.
I wonder if this is a constraint that all AI has, or if it is just an artificial limitation placed on a dumbed down version meant for aiding in search results...
But now that we switched over to looking at individual code blocks, and it's done a VERY good job of documenting them without being needlessly verbose while doing it, I've found that certain questions I was going to ask it were already answered in the small description of the new code block I'm adding.
Sometimes it is way off base. Sometimes it's as if it were reading my mind about what I wanted to do.
But I have to pay attention or I won't know and I'll just be changing code that I would have known wasn't going to fix the problem or would cause more if I took the time to read the AI's description of the problem, what the fix is, an look at the exact code blocks I am to replace and with what.
It seems to be REALLY good at focusing on one module at a time, and I've gotten it to look at the same module for old redundant code that is getting in our way 5 separate turns before it ran out of things for us to fix there (regarding what our aims currently are).
But when we go back to looking at fixing a problem it still hasn't been able to fix in the overall flow of the code, it will actively cite a problem, and give a fix to that problem that doesn't at all remove the legacy code issue in that original code that it just presented. It's as if it forgot we were doing that completely, and it's back to trying to fix our execution code without remembering at all the current reason why it's actually broke in the first place because it's trying to preserve the old and bad code.
I guess we're still in code-overhaul phase. I'm just glad I have a much better intuition about the code than I did a few days ago and I have been able to catch it on several occasions either proposing a bad fix, or misunderstanding/forgetting what an actual goal or the state of the real world testing environment by swapping things incorrectly.
I'm fairly confident this time once we have everything working here, it should also work anywhere else as well. And if that's not entirely true, it should only take a few tweaks to fix it the next time instead of doing a virtual tear-down and rebuild of everything.
Not much worse than that feeling of thinking you've finally fixed the code, only to realize that it doesn't work at all somewhere else. It's tricked me now 3 times into thinking that was the case, but now that I know how much it lies to me and tries to cheat, I know what I've got to do to counteract that.
And part of that really has been to learn more about the code itself.
This can't be done if it has the memory of a goldfish and I have the eyes of a wombat and the gullibility of a sheep.
--------------------------------------------------
Those who dance always seem crazy to those who can't hear the music.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Nice... Almost there. Got the execution code running again, but still have some mismatched pointers and old code in the way of getting it running right.
The AI seems to think we can split up the remaining code after fixing this all to the 3 separate modules in about 10 minutes.
Part of me actually believes that this time.
Not 10 minutes. But maybe an hour.
We just forced out a whole lot of redundant processing and didn't break a single working function while doing it. We're making the code so dynamic/agnostic that I think it would be extremely easy for me to eventually make this a program whose scope and purpose does far more than I initially intended for it to do.
I just can't wait until we finally get it all stripped down and elegant and working again so I can start adding more features, the right way this time.
ETA:
We did it!
At least, I think we did it.
Finally got the "perfect" run, front to back, with all the data giving the exact outputs I desire. And I don't think that there is any more cheating going on under the hood that will make this fail on us in a different environment.
I took a little time out to beautify the output graphs, but now that that's done I'm going to go back and run it on other environments to see if there was any cheating before tightening up our aggressiveness and testing in new environments before splitting up our code further into a few new modules.
Awesome.
--------------------------------------------------
Those who dance always seem crazy to those who can't hear the music.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME