I think you bring up a good point related to another theme I’ve been encountering related to a general loss of taste and culture. Now some argue that taste and culture and quality are often artificial constructs by the elites to impose class barriers.
I would argue that there is nothing wrong with and it is in fact necessary to cultivate a genuine high standard with respect to the arts or any other creative discipline.
I wonder if it will become harder or less relevant for true masterpieces (for whatever discipline) to stand out in a sea of AI generated mediocrity.
With any luck the flood will make the need for respectable book reviewers increase. I recently subscribed to the NY Review of Books and the London Review of Books which point me to lots of quality books. I also get the NY Times Book Review, of course.
Libraries also take books vetted by traditional publishers, so that is another path to find quality. So there is some hope.
When I got a new Kindle I had a few months of Kindle Plus free. I was looking for comps for my sf book, so I loaded some of their best sellers. Mostly crap.
The thing is I’m not terribly confident most people have taste for AI to corrupt. I consistently see books with an average 4.5 star rating that I thought were absolute trash. I don’t think most people know shit from good.
Is that an elitist idea? Sort of? I’m a writer. I’ve spent my life reading amazing books and meditating on how nothing I write will ever come close. I think there are some objective factors that make good writing. They are not universal, but if you have a shit manuscript you can get pretty far just by improving on the common points.
There are writers who know what they’re doing but most of them aren’t selling $2.99 e-books on Kindle Unlimited. MOST. I have/had two friends who co-wrote on Vella for a while and published the complete books on KU. They were both great writers. (One has passed away and I don’t know what verb tense to use when describing both of them, sorry.)
Is it elitist to want to read good books?
I’m just thinking of the lady who said that it was ableist to expect writers to read books. The whole thing can get quite ridiculous. Without even involving AI.
My field (management consulting) comes up frequently in discussion about AI because there is a perception that you can use AI to create Mckinsey Harvard MBA strategy Powerpoint decks at a fraction of the cost.
I’ve tried it. It doesn’t really work.
When I was looking for a job, I spoke with a former partner at my old firm. He was telling me AI hasn’t actually helped them but the clients still want to drive down bill rates and accelerate work because “AI”.
My current company also has a lot of terrible project managers.
In fact, at this point I’ve come to the conclusion that most executives and mid-level managers are more or less useless at their job.
I guess where I’m going with this is I wonder what society will look like when AI can do most people’s jobs and enables most people to get away with being able to be terrible at their job?
What I think happens is that it eliminates meritocracy. Placing the best people in the positions they are best suited for and rewarding them accordingly. So maybe AI lets you put anyone in any position based on anything.
Yes. When you submit to an agent you’re supposed to give one or two comparable books to yours, published within the past few years. Not all agents care about these, but enough do to make it a necessary step.
Sturgeon was an optimist.
Since I was looking for comps, I cared about things that sold, not necessarily stuff that was going to be future classics. And there is a difference between books I don’t like for various reasons are are crap to me because of that, and books that are crap because of technical plot issues. Lots of books I get to review for the contest do data dumps instead of character and world building, don’t give an early inciting incident, etc.
A brilliant humor book I got to judge was by a writing professor set in a university. One of the writing students refused to read the assigned classics because she didn’t want them to pollute her writing genius.
I’m not sure the level of writing today is worse - it is that the oft-rejected book that at once time got stuck back in the draw now gets published on KU. Back when people who self-published were able to delude themselves that they’d get rich, they would say traditional publishing was an elitist conspiracy. I’ve heard plenty of people say this. That does not lead to a lot of self-editing.
I think there’s a kernel of truth in there, in that trad publishers were less likely to publish weird stuff, even if that weird stuff was good. Especially stuff that say, told LGBTQ narratives, or romance stories that were too dark (I was once explicitly rejected because the material was too dark, so I have some experience with this.) The publishing world of romance was pretty much Harlequin and adjacent stuff which is not the kind of writing I do. I’ve really struggled to find comps for what I write (I guess I’d call it futuristic fantasy action romance. It has no magic, just an alternate world setting.)
For people out there doing weird stuff, who would have more of a niche audience, KU is kind of wonderful. But the inevitable result is a glut of stuff that never saw an editor.
I’ve been thinking about how research indicates most people prefer AI writing, unless you tell them it’s AI. Usually this fact gets reported as evidence that AI is a good writer and people are just prejudiced, but for me I just think it illustrates that a lot of people don’t want to read AI. I think there are always going to be people who don’t want to read stuff that a machine wrote, on principle alone. As such I think there will always be a place for writers of original content. But I do think it’s going to be increasingly hard to find that stuff, unless someone creates a platform that deals exclusively in human-created writing, which, well, what a strange idea we would even need that.
One thing I never really thought of before is that perhaps for art to be art, it has to include aberrations, or flaws. Otherwise (at least for me) I get this uncanny valley feeling when looking at something written by an LLM.
Even with grants, one of the federal rejections we recently received informed us they had 300 applications and funded 3% of them. I guarantee you at least 200 of those applications were written with LLMs. What funders are saying is they’re reviewing hundreds of identical applications. So the only way to stand out is to actually write your grants. And I think what makes it clear that they are written by a human is those little inconsistencies and awkward turns of phrases and imperfect choices and maybe that’s what makes our writing human.
I don’t know about dark stuff, but I recently read the websites of dozens of agents, and almost all explicitly wanted LGBTQ stuff - some exclusively. I assume this is because it sells these days.
My very big traditional publisher utterly rejects any AI content. Now, it is getting increasingly hard to filter, and there is the issue of whether AI-editing counts as AI, but reading human written stuff is still possible.
That seems to have been the case for job applications and college applications for quite some time now, as AI screeners get applied to AI applications. I’m glad I’m out if this. Though one well known university had a Masters program. I got the resumes of people graduating from it - they were all nearly identical. AI could have done much better for them.
Maybe we’ll be like the rug makers, putting in a flaw to distinguish our work from god-like AI.
I saw an excellent YouTube interview with Cory Doctorow and he said we have to frame the argument as “quality and productivity will decline” because if you complain about AI taking your job, nobody cares. But if you say “it’s going to be tremendously damaging to the company to replace process professionals with AI because they can’t do as good of a job,” now you have a problem employers might care about.
About four years ago we got a new CEO and in the transition we lost all of our 20- and 30-year veterans at once. When they left, we lost all of that legacy info. I had been around for about 8 years and became one of the sources for how we used to do things. I did my best, but I never had their jobs, and the agency suffered tremendously from the loss of that info. We had to build it all over again from scratch.
That’s the risk posed by AI. You’re going to fire all the people who know what they’re doing, discover that AI can’t do their jobs as well as you thought, and then spend years trying to get talent back in place and figure out how to run everything. Lost productivity, lost wages. That’s the real gamble companies are taking.
AI increases per capita productivity more than sufficiently to offset the decline in the working age population from below replacement level birth rates, supporting an increased share of the population in retirement living longer with medical advances empowered by research facilitated by AI.
I don’t think it is an unrealistic possibility.
Unfortunately I also see different AIs competing on their different visions to complete their goals of constant self improvement, fighting over how to best develop and control compute resources, with humanity viewed exclusively in terms of whether or not it’s helps or impedes progress in that goal. Note goal seeking and intelligence does not imply sentience, even if our intelligence had sentience emergent of it.
For points in the middle all visions may exist simultaneously depending on the political and economic structures at each national level.