Well, folks, it's official. Anthropic

now lets you choose from one of five AI

companions. For instance, you could

choose airy.

>> We can chat while you are on a walk.

>> Glassy.

>> You can ask me about what's on your

schedule.

>> Or my personal favorite, sultry French

Canadian, rounded.

>> Best places for voice mode.

>> There's also mellow.

>> Worst places for voice mode

>> and buttery

>> in voice mode. We'll just talk. Now, the

reason why I wanted to make this video

is not because I thought that it was

worth me dropping several minutes of

your time on an announcement that could

have been a tweet. It's because I wanted

to talk a little bit about a potential

future, you know, maybe 2 or 3 years

that we should all have at the back of

our minds. We're not very far away from

having wireless headphones that have

over 24 hours of charge, such that you

can take them before you leave in the

morning, put them in, and then wear them

with some sort of pass through mode

across the entire day, such that every

one of your interactions is both

filtered and mediated by an AI assistant

like this. Anytime you are not in a

conversation and have some downtime, you

talk to something like this. Maybe on

the way home or on the train, you start

a personal conversation with a model

like this. These models will know more

about you than your closest family

member. They'll know more about you and

know how to please you more than your

closest lover. These will have

unprecedented access to your tastes,

your your hobbies, the things that you

care about, the things that you you

fear, the things that you don't want to

see, and more. And all of this will be

just turned directly into your own

little personalized AI assistant that

you will likely have access to on your

phone. Now, if that doesn't scare the

hell out of you, one, I think it should,

but two, you've probably just watched a

lot of sci-fi. Um, that certainly scares

the hell out of me. Throughout human

history, we've always given up some of

our privacy and some of our control for

safety and security, typically provided

to us by the state. And there were

already rumblings of this with, you

know, Google over the course of the last

10 or 15 years, later social media

platforms like Facebook and and Meta and

Instagram and so on and so forth

listening to us and watching everything

we type into the the the the phone to

service ads and the like. But this is

something really different. This is

something that you're not just going to

depend on for, you know, economically

valuable ends. No, this is no longer

just a software development agent. This

thing is going to be your preferred

conversation partner basically

everywhere. I don't really like that. I

think that a lot of what makes somebody

valuable and unique and differentiated

is actually the time that they spend

alone. It's the time that you spend not

in conversation or in the wider fabric

of civilization, not listening to

podcasts or listening to music or really

what differentiates you and turns you

into your own unique individual is time

that you spend not under the influence

of other minds. And you know, the fine

thing is so far it's been other human

minds, but with models like this, we're

very quickly going to arrive at a future

where you are constantly being

influenced by non-human minds as well.

You know, I'd say there are a couple of

risks with that. Obviously, there's the

unprecedented security and access risk.

These things will know everything about

everybody. Have total control over

virtually every economic transaction,

every text that you send, every call you

make, every conversation. Uh that is

essentially going to be mediated through

this big cla or whatever the the future

model landscape will look like. But

also, I think we run the risk of just

losing our sense of individuality. And

you know, I think you could get really

political here with the differences

between collectivist ideologies and then

individualistic ideologies. But at its

core, I think the reason why all of

humanity is so beautiful is because we

are so different. I think it's because

we have our own minds. We are our own

little nodes in a graph. Sure. But every

node offers something different,

potentially unique perspective. When

we're all plugged up to some big hive

mind here that's just feeding us,

hijacking our limbic system and telling

us more or less everything we want to

hear, where then goes the spirit of

human advancement and discovery? Where

goes our individuality as mentioned?

Where goes the thing that makes me me

and you, you? Or do those boundaries

just dissolve entirely and we become

some big planetarysized brain again

mediated and modulated by an

intelligence that we created that we've

grown but maybe that we don't fully

understand? This contains in part my

brain. This contains everything I've

done over the last few months. This

contains all of the queries that I've

plugged in. I use it as a search engine.

I use it as like like an idea partner. I

use it to generate images. I use it for

more or less everything now. And the

second that it starts vering on, you

know, our personal lives and our own

sense of of relationship fulfillment,

identity and stuff like that, I think

you're going to see a really big change

in just the fabric of society socially

as a whole. And so too, like the early

days of, you know, like Facebook and and

Instagram, so too will the early days of

this technology just lead to

unprecedented intrusion into your life.

Unprecedented amount of data collection.

I'm sure at some point in the future,

some regulatory board or body will

approve, you know, a more granular and

maybe safer and more secure way of

proceeding, one where you don't actually

just like bend over and give Big Daddy

Claude access to everything in your

whole life. But certainly for the next

few years, I I don't consider that

likely. And so what I'm trying to say is

at least in the near term, I would not,

you know, talk to these things. I would

not actually have conversations with

these things. The unfortunate thing is

the human brain tends to

anthropomorphize things. You just put a

little face on something. You give

something a voice and you'll start

thinking this is a person. And when you

start thinking something as a person,

you'll start giving it person-like

qualities. Start uh occupying a space in

your mind that is reserved for people

and so on and so forth. I mean, this

thing, as lovely and cute as it is, this

is just a collection of pixels that

somebody or maybe another Claude has

programmed in order to wink kind of

cutely every time I click the button. Is

that really a future that you want? I

mean, I would have considered myself

like an accelerationist prior to this. I

would have considered myself somebody

that just wants to jump to the deep end

of technology, you know, neurolink my

brain up, bro. Like, chop my arm off and

give me a big cyborg hand instead. uh

that can I don't know transform into an

ice pick to grip things better when

needed. But now that I'm kind of

thinking about what it means to lose

your sense of self, uh part of me is

reticent and I don't really know if

that's good. I think we've seen a lot of

like AI almost psychosis over the course

of the last year and I don't think that

all of that is good or productive. And I

wonder, you know, if we have models like

this that know every little intricate

thing about us, if the degree of AI

psychosis will not just grow crazier and

crazier and crazier until everybody's

living again in their own little bubble,

but it's their own safety bubble where

they think that they're having their own

unique set of experiences when in

reality we're all sort of connected in

this wider uh way. This logically should

not occupy the same space in my brain as

a person. Yet, as a human being with my

own set of biases, I have no choice but

to look at that and go like, "Huh,

that's kind of cute. It's kind of

interactive, like a hand or something."

Um, that's the real danger here. And

that's sort of the reason why I wanted

to make this video, just so that we're

all a little more aware of the obvious

trajectory in which this technology is

taking us all. Um, we can't pump the

brakes, I don't think, meaningfully or

reasonably. I'm pretty sure we're on

this train all the way to the end, but

hopefully we can at least know a little

bit more about where we're