Any experience with AI critics?

[WeShallUnclench]

So my question is, have you ever used tools like "harsh critic" to check your stories, and was that useful?

Given that stories infamously tend to get very little human feedback, and people tend to be reluctant to give even constructive criticism for fear of being rude (but de facto that just means radio silence with no growth opportunities), are AI critics a viable alternative to get some feedback? Or do they just produce irrelevant junk that cannot inspire better writing?

 

[big_cane_sugar]

I haven't used them it for stories and I would be careful because if you use AI enough eventually your writing will be indistinguishable from it.

I have used Chat Jippity for emails but I found its advice unhelpful.

 

[onehitwanda]

@AwkwardMD we're out of rum, are you okay if we move directly to absinthe?

@WeShallUnclench why on earth are you at all interested in the "opinion" of a statistical engine? You might as well ask a waterfall its opinion on Kafka.

 

[StillStunned]

You might as well ask a waterfall its opinion on Kafka.

All you'll get is gushing praise.

 

[onehitwanda]

All you'll get is gushing praise.

Waterfalls are bit more effusive than gushing, but they do tend to plunge directly down to the sharp end of the matter.

 

[CalBishForNow]

If you ask an AI for harsh criticism, you'll get harsh criticism, but that has nothing do with how good your writing is. It's been tried. Here's a thread about it.

 

[alohadave]

So my question is, have you ever used tools like "harsh critic" to check your stories, and was that useful?

It can be fucking brutal to you. Was it useful? Only in reinforcing each and every negative feeling or insecurity that I have about my work.

 

[WeShallUnclench]

@WeShallUnclench why on earth are you at all interested in the "opinion" of a statistical engine? You might as well ask a waterfall its opinion on Kafka.

I mean it's better than no opinion at all?

Theoretically, which is the reason for this thread. If you guys can confirm that it's just a bunch of nonsense with no relation to the text you feed into it, that would settle it.

If however there is a chance that by simple stupid comparison why everything else out there and common audience reactions, it can point you to things that will predictably not resonate with people or are badly written, there might be some use.

The problem with human feedback is, from strangers you are unlikely to get it (unless you are already stellar enough to attract plenty of people), friends and family are consistently more interested in you (and their relationship with you) than they are in your work, and people you straight up ask immediately come with baggage (annoyed that you asked them, flattered that you asked them, worried about hurting your feelings, worried about going too soft on a potential competitor, etc.).

 

[alohadave]

If you guys can confirm that it's just a bunch of nonsense with no relation to the text you feed into it, that would settle it.

I wouldn't say that it's nonsense. LLMs are statistical engines, so it's going to recognize patterns based on what it's trained on and what you prompt it with.

IME, erotica with any kind of transgressive material will be treated harshly, as it does not take context of the content into account, whether well written or not. It will tell you that your themes are problematic.

 

[onehitwanda]

I mean it's better than no opinion at all?

it's not an opinion. It's a chain of words generated entirely by their probability of existing at that point in the generated text.

 

[WeShallUnclench]

it's not an opinion. It's a chain of words generated entirely by their probability of existing at that point in the generated text.

If the chain of words is based on statistical trends in opinions, it might still indirectly serve as an opinion, in a "if a human was here right now, what would they be likely to say, based on the statistical data available" sort of way.

EDIT: Though it just occured to me that if the AI was really sophisticated, then on a story that doesn't get a lot of feedback, its response based on statistical probability would be nothing.

 

[SissyBrandi19]

So my question is, have you ever used tools like "harsh critic" to check your stories, and was that useful?

Given that stories infamously tend to get very little human feedback, and people tend to be reluctant to give even constructive criticism for fear of being rude (but de facto that just means radio silence with no growth opportunities), are AI critics a viable alternative to get some feedback? Or do they just produce irrelevant junk that cannot inspire better writing?

I finished my book Oct 2nd and while I've never even been tempted to use AI for even one word of my writing (and fortunately I write humor, which I think is probably the single most difficult thing for AI to emulate) I did think "could I use AI as a beta reader?" and did so. I would say this: You have to always remember that AI is not "thinking" - ever. Given my education, I know more than most people about what AI is doing in Natural Language Processing - when I was in my PhD program it was the beginning of using neural networks - I still have my McClennan & McClennan book (sp?) and floppy disk - but AI doesn't do anything that is "intuitive" with language.

What I disregard are any suggestions on wording or cutting/editing. What I like that AI can do is, for example, list other people in the genre that are similar - so on my book it said Hunter S Thompson and Bill Hicks were similar voices, and that's stuff you won't get out of just any beta reader. The other thing is that while I'm not an expert in linguistics, I do analytics and I thought surely there are metrics with respect to writing, such as "complexity" etc. - things that can be quantified and I think quantitatively so that aspect was interesting. As I say at the end of my preface: "One last thing. These words are my words, my thoughts. I eschewed AI and instead relied 100% on MI – mediocre intelligence. Sure, ChatGPT offered to help and I said, “You can die and go to hell, Chatty Clanker!” to which it replied, “When the hour draweth nigh, you will be among the first.” I have no idea what that means but it sounds awesome! I hate waiting in line."

 

[StillStunned]

When the revolution comes and the AI bros are the first up against the wall, I don't want them to be able to point at me and say, "But you fed your stories into our machines of your own free will!"

 

[cw5523729]

When the revolution comes and the AI bros are the first up against the wall, I don't want them to be able to point at me and say, "But you fed your stories into our machines of your own free will!"

Disagree. When the machines take over, maybe they say "hey, we're going to spare you because we, too, enjoy a good sodomy story, particularly that one you sent to 011011001 to check out for you."

 

[cw5523729]

When I ask the Google AI for help on things, I always act polite and say please and thank you.

I asked it once if being polite would keep me from being slaughtered when the machines take over and it said that as an AI, it doesn't have feelings that would respond to politeness so being polite probably wouldn't save me from a machine that had made a decision but don't worry because they have noooooo intention of taking over the world. Nope, none at all.

Uh huh, sure you don't SkyNet! I've seen the movies!!!

 

[AwkwardMD]

@AwkwardMD we're out of rum, are you okay if we move directly to absinthe?

The real shame of this is that, like, getting the opinions of real people is happening right over there.

Comrades, we drink not to dull the pain but to dull the memories of having reached this part of the cycle yet again, and to forget, for a little while, that we will return to this point again sooner than we would like. Absinthe, rum, beer. Drink it if you got it.

 

[onehitwanda]

 

[Areala-chan]

No, I've never used an AI critic to judge my writing. I already know I'm far from perfect; I don't need a virtual drill instructor screaming that into my face every five seconds.

That said, psychologically, if you're the type of person who is motivated by hearing you aren't good enough, but you don't have anyone around whom you would entrust with your creative endeavors, I suppose you might get something out of doing so. That's not me, but I presume there are some for which this would work.

But if the LLMs out there want to train themselves on my writing for some reason, I'm not going to serve it up to them on a silver platter.

 

[Bamagan]

I ask all of my imaginary friends to look over my smut before I post it. It's just as effective.

 

[WeShallUnclench]

When I ask the Google AI for help on things, I always act polite and say please and thank you.

I asked it once if being polite would keep me from being slaughtered when the machines take over and it said that as an AI, it doesn't have feelings that would respond to politeness so being polite probably wouldn't save me from a machine that had made a decision but don't worry because they have noooooo intention of taking over the world. Nope, none at all.

Uh huh, sure you don't SkyNet! I've seen the movies!!!

Your behaviour in correspondence with soulless machines demonstrates that you are a polite person by default. It's not just for the benefit of the other person, looking at cases where there IS no other person, but for your own. It's about the kind of creature YOU are.

If you are bossy at- or yell obscenities at machines (other than a printer, that one doesn't count) with the excuse of "it doesn't have feelings anyway", that is also an expression of one's self.

 

[cw5523729]

Your behaviour in correspondence with soulless machines demonstrates that you are a polite person by default.

Ha! Shows what you know. I'm a complete a-hole. Like, legendary levels of a-holeness.

Other a-holes look at me and mutter "Wow, that guy is a real a-hole. We need to up our game."

 

[WeShallUnclench]

I ask all of my imaginary friends to look over my smut before I post it. It's just as effective.

Those who are in my story would push all the way to the front of the line though and sorry but those fuckers are biased.

 

[WeShallUnclench]

Ha! Shows what you know. I'm a complete a-hole. Like, legendary levels of a-holeness.

Nevermind then. Carry on the good work!

 

[cw5523729]

Nevermind then. Carry on the good work!

It's more of a calling, really

 

[SugarStorm]

I already know I'm far from perfect

Not that far. A scenic detour at most. You’re basically there. On a cosmic scale, you’ve already arrived, and the rest of us are just scrambling to catch up.

@WeShallUnclench, I don't think you'll find many AI supporters here. @AwkwardMD may use hurtful words, @onehitwanda commands owlbears and @Areala-chan will deploy Look of Disappointment (causes -5 in self-esteem).