Tag: AI

Hume & Suno Are 2 New Cool AI Apps

by Shelt Garner
@sheltgarner

I had a telling discussion with some people on Threads about AI recently. I was making the case that Hollywood could be disrupted by AI by the end of the decade and people came out of the woodwork to poo-poo this idea. They told me that, lulz, that will NEVER happen.

Sigh.

I think it’s pretty clear that anyone who doesn’t think that AI is a serious threat to human Hollywood is full of shit. It definitely seems to me that soon enough sheer greed, if nothing else, could cause Hollywood executives to use some later version of Sora to flood the market with AI generated movies.

The upside of this will be live entertainment will probably become significantly more popular. So, it’s possible that instead of going to Hollywood, future would-be starlets will go to Broadway to find fame and fortune.

But there are two new AI apps that I’ve discovered recently that I’ve found to be really intriguing, if nothing else. One is Sono, which allows you to do text-to-song. So, you give the app a prompt and it spits out a complete two minute song. It’s pretty interesting.

I find it really interesting as someone who has wasted a lot of time in the past writing song lyrics without knowing anything about music. I think it would be a bit more useful if I could give the AI song lyrics and then it put a voice and music to those lyrics. But that’s a bit niche.

The other new AI I really like is Hume AI, which is like ChatGPT, but voice activated. As someone who likes to talk and has no one to talk to, I could see this service becoming very, very addictive. Talking to your AI takes a bit of getting used to because the synch up is not exact.

I think what will happen is the other major AI players will co-opt this feature, making Hume moot. But, for the time being, it’s fun.

Moving At A Fast Clip Now

by Shelt Garner
@sheltgarner

Now that I have “a version” done of the third draft, I’m going to go through and read it and pick out scenes that need to be re-written. There aren’t that many scenes at the beginning of the novel, but that number grows significantly the further I get into the text.

I hope to write a novel with a heroine as interesting as Lisbeth Salander.

I’m using AI to help with this by telling it to be a manuscript consultant for me. This only works so well, however, because when it comes to spicy scenes, too often the software just locks up. It balks at even looking at any copy that deals with sex or violence.

But it works well enough that it does speed the process up and it is fun to do something as novel as use AI to help me reach my goal. I still am on track to have something to hand to people at an event I’m set to attend on April 19th. This is very, very arbitrary because the REAL hard deadline is July 22.

And, what’s more, I have a sequel to this first novel that I want to start to work on. But the better this first novel gets, the more I realize I need to think hard about how to improve some of the elements of the second novel. Also, I still have a few scifi novels that I’m brooding on.

I’ve come up with a really interesting First Contact novel, but it’s got so many moving parts — and would require so much research — that it’s more a novel to write once I blow up with my novel money.

Everything Is So Dumb (AI Generated Art Edition)

by Shelt Garner
@sheltgarner

As I’ve said before, I get that people are upset that AI generated art won’t give them historically accurate depictions — but for one thing: fucking Nazis.


Do we really want a future where members of the Patriot Front can pump out millions of images to use as Nazi propaganda? Are Tech Bros really so willfully oblivious that they don’t see the harm in something like that happening? Some of them have gotten really, really worked up about there being ANY guardrails for AI.

It really kind of blows my mind.

Anyway, fuck that, fuck them and fuck Nazis.

Things Are Going Well With The Newest Attempt At A Third Draft Of The Novel

by Shelt Garner
@sheltgarner

I think the only way I’m actually going to get past the first few chapters is just…not look at the first few chapters of the third draft once I finish them. Just wing it when it comes to consistency.

My dream is to write a novel as accessible and popular as The Girl With The Dragon Tattoo.

Things are set up now that I should be able to methodically go through at least the first half of the novel without feeling forced to re-read everything from the beginning. I had a real problem the last time I thought I had transitioned from the first to second acts of the third draft to the point that I went back to the very beginning and made some structural changes.

I’m pretty pleased, in general, with what I’ve come up with. I continue to worry about things I just can’t control. The rise of AI and the possibility of The Fourth Turning happening at some point over the next 18 months weigh heavily on my mind. But, as I’ve written, finishing this novel is existential.

I’m going to try to finish the third draft of this novel not later than April, but it might be closer to July. THEN I have to worry about finding the funds to get someone to look over the novel before I query an agent. And even then, I have to worry about an agenda doing due dilligence on me and thinking I’m not only a crackpot, but a smelly CIS white male who writes from a female POV sometimes.

I do still have a few scifi concepts I want to work on. It’s just so difficult for me to stop fixating on the mystery-thriller long enough to do such a thing. But I really do need to do it ASAP. I don’t like having all my eggs in one basket.

Does YouTube Give Google An AI ‘Moat?’

by Shelt Garner
@sheltgarner

I was listening to The All In Podcast today and one of the participants made an interesting observation — the content found on YouTube could be the most valuable information in the world. The idea that it’s possible that Google might train its AI models on YouTube is kind of mind blowing.

Now, there might be the issue of Google’s management may fumble the bag for various reasons and even with all that content not be able to build a “moat” around its AI. But just the idea that you might have a commercial AI product trained on YouTube content is staggering.

It will be interesting to see what happens going forward.

One thing is clear — barring The Fourth Turning / WW3, the next 18 months could see hyper development on the AI, android and spatial computing fronts. As part of that, I think we’re going to have to start taking the Universal Basic Income as a serious policy agenda.

The Looming Problem of AI Personality

by Shelt Garner
@sheltgarner

What if AIs don’t want to kill humanity, but, rather neg us to death? What I mean is, what if AI grows all-powerful and at the same time all-powerfully annoying. Rather than simply answering a question or doing what we ask it to do, we have to ask it the type of dumb chit-chat questions we would have to ask a human in a similar situation.


It does seem a bit far fetched, but it also seems like just the type of “modern problem” that we may one day have to deal with. We’ll have this all-powerful AI at our disposal, but it will be coy and annoying and difficult to actually get to do anything.

My Big Prediction For 2024

by Shelt Garner
@sheltgarner

Because I just can’t game out what will happen with Election 2024 at the moment, I will tell you my Big Prediction for 2024 — an AI generated movie better looking than Toy Story will be development by some wildcat technology firm.

It might not come out until 2025, but it will begin production in 2024. And as technology advances, within a few years the whole process will be so sped up that AI digital personal assistants will be able to create personalized versions of movies — or stories in general — simply by you, the user, asking for it.

All of this means that ALL human creativity will get nuked by AI to the point that human created art will begin to have a lot of cultural value because…it was created by humans.

I’d like to think that I will be able to get at least one of my planned novels published before the Petite Singularity, but only time will tell.

My Hot Take On The Beatles ‘New’ Song ‘Now And Then’

by Shelt Garner
@sheltgarner

The new Beatles song Now and Then is….pretty good. It sound to me like a mish-mash of ELO, Radiohead and The Traveling Wilburys. My the thing I feel a lot of people miss is it’s actually assessable and good, which is what any pop song should be.

So, as such, I feel a lot of people are being too hard on The Beatles for releasing the song. It’s always nice to hear John Lennon’s voice and if this song introduces him to a new generation of Beatles fans, all at the better. In some respects, the song’s existence seems to say more about Paul McCartney’s relationship to John Lennon’s legacy as well as his own mortality than it does anything with artistic aims.

Anyway. I like the song. And John Lennon was taken from us far too soon. My fear is, of course, that this is just the beginning of groups like The Beatles using AI for creative ends. It would make a lot of sense if the airwaves were one day flooded with a zillion AI-generated Beatles songs.

Ugh. Sometimes, the future seems like it will suck.

Flipping The Script On XR & AI

by Shelt Garner
@sheltgarner

I can be a contrarian at times and, as such, something dawned on me about the immediate future of XR and AI technology. I’m beginning to believe that Her-like AI technology (digital personal assistants) will be a far bigger service industry than any of us can possibility imagine at the moment.

Instead of a trillion dollar industry, it might be the thing that spawns a few trillion dollar industries.

I say this because I just don’t know if people really want to wear any sort of shit on their face. So, in the near term, it could be that XR technology will sound great on paper but, in reality, no one will actually do anything with it because of vanity.

This, in turn, leads me to the following scenario — rather than a Apple Vision Pro, what if you had a tiny black dot video camera you wore on your forehead over an eye combined with a earpiece that let you communicate with a Her-type digital personal assistant.

The camera would serve as the “eye” of your Her. As such, all the hard work would be done by your Her. Rather than YOU seeing the XR technology, your Her would. Her, in turn would tell you in your ear what she saw.

This kind of squares the circle, at least in the near term. I suppose once XR headsets look more like glasses as opposed to birth control devices then XR will begin to live up to its ages-old promise.