It's all about jokes, funny pics...

I changed my mind a bit and fine Claude Code quite useful, if you go in small steps and review the coding.
Next level you'll gain is, when you learn reviewing always takes longer than writing.

When I write my posts, the first sketch is written within a few minutes.
But since I respect my readers I want to give them something good to read.
So I review it.
To review it afterwards - correct all typos, misspelled words, wrong words, wrong grammar, throw out redundancy, improve the wording and undestanding, split and shorten sentences, re-formulate the one or the other thing... - takes at least ten times a first sketch needs.

If I do this on something some AI chatbot barfed out, it saves me most of those few minutes to get to the first sketch.
For the price it takes me at least twenty times to review and modify a text not written by me, just to make it something like I might have could said, so dealing with somebody else's thoughts, in this case not even thoughts but some machine's patterns overflowing with redundant babbling, while keeping my own thoughts about what I actually wanted to say in the first place at the same time. 😵‍💫
This ain't neither easement, nor improving the fun.
Except, you don't give a sh1t on the result's quality and don't care about your readers. But if you don't care about your readers, better don't write at all in the first place. Writing is something you do for others. It's immodesty to throw some bad text at your readers without considering how they gonna take it.
Writing is for reading. Writing just used as for it's written is like a WOM, a Write Only Memory. Pointless.

And with software's source code this is even worse.
You may get something working from it quickly and easily, yes, but... - there is a large but, and I don't want the 5,371st "AI" thread.
Short: When your quality standards are that low that reviewing takes same or less than producing the first sketch, then some AI chatbot is very useful on producing stuff. That I can grant, sure.
 
tom.png

"Hey, ChatGPT, I read-------"Ha, ha, ha, what a load!
that AI will erase-----------How shall this work?"
humankind."
 
To review it afterwards - correct all typos, misspelled words, wrong words, wrong grammar, throw out redundancy, improve the wording and undestanding, split and shorten sentences, re-formulate the one or the other thing... - takes at least ten times a first sketch needs.

Nawwwww... I don't do that :)

You should be able to get everything right on "the first take"!

I mean TV shows used to be filmed before a LIVE audience !
 
You should be able to get everything right on "the first take"!
I don't know, if I should be able of that. But since I'm not I find it better to review than wild unconsidered typing.
And looking at many other posts I doubt everybody else but me is capable of getting everything right in the first take, even if they think so. IMO many things I see posted here better be reviewed, often even being considered twice if posted at all. Reviewing requires the ability of self-criticism. But there are some who don't even know what criticism actually is, others who suffer of hubris, and some combine both pretty well.
I am fully aware of the mainsteam does not match my quality standards, since I refuse to read, listen to, watch, drink, eat, buy, use and vote for 💩. 😁
I mean TV shows used to be filmed before a LIVE audience !
Every audience in every TV show is carefully selected, briefed, and warmed up before a show. They get signs when to applaude in which way, when to laugh, when to stop etc. TV always is a show, no matter what they air, including the news.
NOTHING on TV ever happens 100% uncontrolled, spontaneously, free and being aired completely unfiltered.
Even if it supposes to seem this way plausible that there was no script and it is uncut live, like all these "reality" or "docu" BS, there always is at least some preliminary briefing, and one editor can pick from several camera views, decides what is actually aired, to prevent airing unwanted things.
 
I don't know, if I should be able of that. But since I'm not I find it better to review than wild unconsidered typing.

I think your point on AI is correct ! (And this might be why they are failing with AI right now)

Coming up with an "idea" (or a "thought") means that one keeps thinking about the idea and improving the idea. As you mentioned correct typos, throw out redundancy, pick pros and cons, evaluate predicted outcomes, etc.

The main problem with AI right now is AI's rush to PICK the first thing "it thinks of" (aka computes) ... and THAT'S IT ! When presented with "thought branches" (A), (B) and (C) -- it immediately picks the best of [A,B,C] and presents (for example (A) ) to the User/UI and then AI completely forgets about all the work it did to compute the result. (Or maybe) AI remembers the result, but doesn't question "itself" again about how it arrived at the result. Of course new "thought choices" (D), (E), and (F) could suddenly present themselves to AI, but too late for AI to consider because AI has already started processing some "spaghetti recipe" or something else.

Sadly AI can look at [A,B,C] AGAIN (even a short time later) -- and suddenly pick "thought branch" (B) -- but the Human is left to scratch their head as to why the AI chose answer: (A) at 12:00pm and (B) at 1:00pm

But that's not how humans work.

You are correct that we are NOT always right the first time. We write a sentence that seems right, and we actually believe the completed sentence "is perfectly right", and wouldn't change a thing ..... until the (NEXT DAY) we look at the same sentence again and go -- "What was I thinking?"

The "computer brain" in the Disney Marvels movies was not always immediately right -- and the "computer brain" would inform the human that it would need to take some time to figure things out. I think the future of AI will be "less immediacy" and "more correctness"... well... at least I hope so :cool:
 
Back
Top