I just hiked up Rattlesnake Gulch, which is about 1100 of climb.
I was not physically capable of this feat earlier this year.
I just wanted to say thank you to everyone who has been kind and supportive on my journey through burnout and chronic illness.
I honestly thought I might not make it at times, but today is the most vigorous hike I've done since I ran Tough Mudder in 2014. That was a different life. A different version of me.
What I want to share with all of you is that if you keep chipping away at whatever problem you have, you might be surprised at how far you can get.
And really, this is what life is all about. Yeah, technology is cool and all, but the point of technology is to live better. I'm looking forward to spending more time in my body and in connection with people.
Hey fam, Suno v6 just dropped and it is pretty darn good.
I know some people get mad when I share my music but like, I don't get that. I'm super pro-AI so if this makes you mad then like... that's just your opinion, dude.
Anyways, stoner/desert/psychedelic rock is one of my favorite genres and I wanted to share this track that I've been working on and improving for a while. It really shows how far this tech has come.
What I want to point out is that the AI-artifacts are *almost* invisible. Human voices still have a slightly scratchy texture sometimes, but the shimmer is pretty much gone.
This is salient because of something I talk about a lot in Post-Labor Economics - most people care about the outcome, not the provenance. I would be super jazzed if there was a band who really sounded like this. But there isn't. So my particular taste is only being satisfied by the machine.
And I think this is indicative of where things are going. Personalized music is here. Yeah, you have to make it yourself for now, but it's only a small hop, skip, and a jump until Suno (or someone) allows you to create radio stations based on seeds and then you have unlimited music for whatever genre, vibe, or need you have. Yes, there are still some things it absolutely cannot do (like traditional indigenous music) but I figure that's only a matter of time.
I'm working on a video presentation about this, which will be very precise about exactly how thin the moat that humans have truly is.
In the meantime, I hope you enjoy this track. It's probably one of the best songs I've ever made. It's based on the indigenous myth of la loba, the wolf woman bone collector.
But yeah, in a few years, I think the economics of art and entertainment will be very, very different from how they have been.
I'll be sharing more of my own original research. I'll be publishing it on my Substack mostly, with reposts here and elsewhere.
The reason is because, well, I've become more confident in my work. I've been iterating on some papers and ideas for a long time now, and I figure it's time to get the information out there.
I have a few hopes/goals for this.
First, I'm hoping it will spur conversation, debate, and get people thinking more in terms of post-labor economics.
Second, I'm hoping it will expand my reach. A good idea is only valuable if the right people know about it.
Third, I'm hoping it will help with the launch of my book and other work. Maybe it will lead to good connections.
Let me know if you have any questions.
Also, to be more precise instead of vague, the first paper will come out shortly, and it is about mounting evidence that the great decoupling is actively happening.
My worst take of all time was doubting the importance of Chain of Thought reasoning.
Strawberry man (the Twitter personality that a lot of people hate) told me I was dead wrong.
Several OpenAI employees actually blocked me over it.
CoT led directly to tool use, and basically every ability that makes Fable and Astra amazing.
Jensen Huang and others are saying AGI has been achieved. My original AGI prediction time nailed the arrival of CoT reasoning in commercial models.
I'm not saying "I was clearly right in hindsight!" but what I will say is that I was directionally correct, and while many people confidently say "clearly LLMs will not achieve AGI or ASI?" I don't agree with them.
This paradigm clearly has a lot of headroom left. It will very obviously carry us well past human capability into the realm of superhuman.
That doesn't mean it will be the last paradigm of machine intelligence. But LLMs plus reasoning is actively surpassing human limits.
That reaction, at the time, was by far my most categorically wrong take of all time. I did not correctly price reasoning at the time and I try to remember that mistake every time I try to gauge a new innovation or paradigm.
With that in mind, I will say don't sleep on persistent multi agent platforms like Grok Bot.
David Shapiro
Let's take quick break from AI.
I just hiked up Rattlesnake Gulch, which is about 1100 of climb.
I was not physically capable of this feat earlier this year.
I just wanted to say thank you to everyone who has been kind and supportive on my journey through burnout and chronic illness.
I honestly thought I might not make it at times, but today is the most vigorous hike I've done since I ran Tough Mudder in 2014. That was a different life. A different version of me.
What I want to share with all of you is that if you keep chipping away at whatever problem you have, you might be surprised at how far you can get.
And really, this is what life is all about. Yeah, technology is cool and all, but the point of technology is to live better. I'm looking forward to spending more time in my body and in connection with people.
Let's build that future together.
7 hours ago | [YT] | 249
View 22 replies
David Shapiro
There are two buttons before you. You must press one.
BLUE BUTTON: Nothing changes
RED BUTTON: Erase AI from existence
Which do you press?
2 days ago | [YT] | 92
View 117 replies
David Shapiro
Hey fam, Suno v6 just dropped and it is pretty darn good.
I know some people get mad when I share my music but like, I don't get that. I'm super pro-AI so if this makes you mad then like... that's just your opinion, dude.
Anyways, stoner/desert/psychedelic rock is one of my favorite genres and I wanted to share this track that I've been working on and improving for a while. It really shows how far this tech has come.
suno.com/s/SNHVXDymrKSbHfTq
What I want to point out is that the AI-artifacts are *almost* invisible. Human voices still have a slightly scratchy texture sometimes, but the shimmer is pretty much gone.
This is salient because of something I talk about a lot in Post-Labor Economics - most people care about the outcome, not the provenance. I would be super jazzed if there was a band who really sounded like this. But there isn't. So my particular taste is only being satisfied by the machine.
And I think this is indicative of where things are going. Personalized music is here. Yeah, you have to make it yourself for now, but it's only a small hop, skip, and a jump until Suno (or someone) allows you to create radio stations based on seeds and then you have unlimited music for whatever genre, vibe, or need you have. Yes, there are still some things it absolutely cannot do (like traditional indigenous music) but I figure that's only a matter of time.
I'm working on a video presentation about this, which will be very precise about exactly how thin the moat that humans have truly is.
In the meantime, I hope you enjoy this track. It's probably one of the best songs I've ever made. It's based on the indigenous myth of la loba, the wolf woman bone collector.
But yeah, in a few years, I think the economics of art and entertainment will be very, very different from how they have been.
3 days ago | [YT] | 105
View 43 replies
David Shapiro
I'll be sharing more of my own original research. I'll be publishing it on my Substack mostly, with reposts here and elsewhere.
The reason is because, well, I've become more confident in my work. I've been iterating on some papers and ideas for a long time now, and I figure it's time to get the information out there.
I have a few hopes/goals for this.
First, I'm hoping it will spur conversation, debate, and get people thinking more in terms of post-labor economics.
Second, I'm hoping it will expand my reach. A good idea is only valuable if the right people know about it.
Third, I'm hoping it will help with the launch of my book and other work. Maybe it will lead to good connections.
Let me know if you have any questions.
Also, to be more precise instead of vague, the first paper will come out shortly, and it is about mounting evidence that the great decoupling is actively happening.
1 week ago | [YT] | 156
View 7 replies
David Shapiro
My worst take of all time was doubting the importance of Chain of Thought reasoning.
Strawberry man (the Twitter personality that a lot of people hate) told me I was dead wrong.
Several OpenAI employees actually blocked me over it.
CoT led directly to tool use, and basically every ability that makes Fable and Astra amazing.
Jensen Huang and others are saying AGI has been achieved. My original AGI prediction time nailed the arrival of CoT reasoning in commercial models.
I'm not saying "I was clearly right in hindsight!" but what I will say is that I was directionally correct, and while many people confidently say "clearly LLMs will not achieve AGI or ASI?" I don't agree with them.
This paradigm clearly has a lot of headroom left. It will very obviously carry us well past human capability into the realm of superhuman.
That doesn't mean it will be the last paradigm of machine intelligence. But LLMs plus reasoning is actively surpassing human limits.
That reaction, at the time, was by far my most categorically wrong take of all time. I did not correctly price reasoning at the time and I try to remember that mistake every time I try to gauge a new innovation or paradigm.
With that in mind, I will say don't sleep on persistent multi agent platforms like Grok Bot.
1 week ago | [YT] | 254
View 25 replies
David Shapiro
Here's my first interview in a while. Talking about Post-Labor Economics.
Hope you enjoy!
1 week ago | [YT] | 14
View 3 replies
David Shapiro
I finally looked up the viral Ed Zitron video and... all his talking points are basically a compression of discourse from 2024-25.
And he seemingly has not updated his priors, beliefs, or observations.
It's like a weird time capsule.
2 weeks ago | [YT] | 182
View 53 replies
David Shapiro
2 weeks ago | [YT] | 107
View 5 replies
David Shapiro
I'll be bringing the slide decks back
What do you want to learn?
2 weeks ago | [YT] | 119
View 39 replies
David Shapiro
I've thought a lot about longevity.
Imagine living 500 years. Or 5000.
Sometimes I think about the existential vertigo that comes with such contemplations.
But then I remember...
We're hardly the same person from one decade to the next. Let alone centuries.
The version of you today will be quite different from the version of you that might need to deal with the edge of eternity.
2 weeks ago | [YT] | 300
View 82 replies
Load more