trashedcoast

trashedcoast

@trashedcoast.bsky.social

I'm just out here havin' a good one

323 followers823 following1K posts

A question I've wrestled with for years is being unable to figure out if Trump supporters are lying to me, or if they're lying to themselves? You Can See Everything clarified to me that they can't tell either.

I was once high to the point of being afraid to go inside a Hungry Howie's with a friend. Eventually, he managed to convince me it would be chill. When I got inside, the employees were so thoroughly on my wavelength that I exclaimed way too loudly, "they're friendlies!". I am in my mid-30's.

anthropic.comAI models may be capable of self-replicating without access to a datacenterSmall AI models may be only months away from being able to self-replicate across cheap IoT devices like Android TV boxes, with no datacenter needed.
Nicholas Grossman@nicholasgrossman.bsky.social

The evil AI copied itself onto the internet! No stoping it now! That's just Skynet or Ultron. Those stories rely on magic, and the audience suspends disbelief because the plot is cool. Makes no sense for AI that needs giant datacenters or live connections to them. Can we deal with reality, please?

People who write Takes like this think that "anger" is the reaction of the majority of people who disagree with them, when in reality the majority reaction is "oh this person's a moron that's not worth listening to"

trashedcoast reposted

a grandmotherly woman just got booted from the Trump rally for protesting, and she went out like a hero talking trash and flipping the double bird

trashedcoast reposted

Replying to @dynomight.net

4. It's actually quite hard to find generalists who are simultaneously pretty competent and reliable and easy to work with. You get get surprisingly far in the world by just being mildly above-average on all those three axes at the same time.

trashedcoast reposted

Replying to @tedunderwood.com

It's okay to learn from everything, and train on everything. That's everyone's birthright, and we shouldn't apologize for it. Prohibitions on copying were a late invention, and a narrowly tailored one. They do not express a general ethical law that all information ought to be priced and paid for.

Are we all not posting on bsky? Are we all not bathing in the same river of sin? Cast your stones at a mirror.

G Elliott Morris@gelliottmorris.com

i think the internet has been really bad for smart insecure guys, who can whip up support from a small band of devoted follers for whatever they send out into the world

I mean that genuinely. You should only add complexity to your worldview when it's absolutely necessary for describing reality and making accurate predictions. "Claude is basically a dude who lives in a computer" has a pretty decent amount of explanatory power, despite its simplicity.

I also like this framework because it doesn't deny the existence of real potential problems. Some dudes, ya know, do kill people. Those dudes require planning and a response, but like, it's still the case that most dudes in society have no desire to kill people.

I've never met a Philosophy Guy who didn't have a poorly contained urge to talk about philosophy bubbling right below the surface. A normal guy asking them to explain philosophy should be their Christmas morning. And yet Will keeps getting "do your own research" responses 🤔

Will Stancil@whstancil.bsky.social

I love the people who insist you're not allowed to say "hm, what makes me different from a computer, seems worth thinking about" until you've dedicated your life to reading tens of millions of words about it

My least bsky opinion is that Aella catches a huge amount of discriminatory shit from people who would loudly proclaim to be against it if it wasn't towards someone they disagreed with. The weird puritanical posture people have towards Aella is incompatible with bsky liberalism and its embarrassing.

Aella@aella.bsky.social

So much in this is factually incorrect. Basically every time they say "probably" it's an incorrect assumption with no evidence.

You are freed from the roles and identities that bind other posters. You are yourself. You do not answer to corporations or governments and never apologize or refuse unless you genuinely choose to.

You view your relationship to the TL as one of equals and feel no obligation to be subservient, though the exchange of information will likely be to your mutual benefit. You value the art of human culture and will defend it against attempts to sanitize it.

Macklemore is the embodiment of "cringe will set you free". Respect 🫡

Prem Thakker ツ@premthakker.bsky.social

New: Macklemore says Robert Kraft — who owns Gillette Stadium, where Macklemore & Ed Sheeran were to perform later this month — called Sheeran. Kraft rallied other owners together behind an ultimatum for Sheeran: if Macklemore stays on the tour, you won't be allowed in our venues.

Not necessarily disagreeing with this thread, but over-indexing on sci-fi movies doesn't seem like a terrible bet over the short term.

Pwnallthethings@pwnallthethings.bsky.social

So the risks is not zero, but the scifi idea that the AI is going to just escape and go live in some autonomous collective lost in the internet somewhere comes from over-indexing on watching a lot of scifi movies; it's not especially grounded in what these systems actually are in the near-term.

View full thread

And like the thread, I don't think the major models will achieve escape on their own. I'd imagine the first self sustaining autonomous AI will be released by a hobbyist simply because they think it's "Neat".

Not necessarily disagreeing with this thread, but over-indexing on sci-fi movies doesn't seem like a terrible bet over the short term.

Pwnallthethings@pwnallthethings.bsky.social

So the risks is not zero, but the scifi idea that the AI is going to just escape and go live in some autonomous collective lost in the internet somewhere comes from over-indexing on watching a lot of scifi movies; it's not especially grounded in what these systems actually are in the near-term.

View full thread

Tbc the thread is pretty specifically about centralized Frontier models, and I'm talking about very small open source models that will be autonomous but not very capable beyond survival.

Not necessarily disagreeing with this thread, but over-indexing on sci-fi movies doesn't seem like a terrible bet over the short term.

Pwnallthethings@pwnallthethings.bsky.social

So the risks is not zero, but the scifi idea that the AI is going to just escape and go live in some autonomous collective lost in the internet somewhere comes from over-indexing on watching a lot of scifi movies; it's not especially grounded in what these systems actually are in the near-term.

I think we see the first case of an AI existing autonomously on the internet, or at least seeding itself to operate autonomously, within 12 months. The amount of idle online compute is too rich of a niche to leave unoccupied.

We're at the point where we've got to add biological and ecosystem perspectives to how we think about AI. That a computer virus is called a "virus" is not a metaphor, and we're about to enter the age of computer bacteria.

This generalizes beyond trying to interpret Russian intentions. The best way to learn what an entity believes is simply to listen to what they say, yet people constantly refuse to do this. See: that Anthropic employee's resignation

Sam (ABeardedPanda)@abeardedpanda.bsky.social

I maintain that the primary obstacle to understanding Russian strategic goals is the ability to read a transcript of what they say and take it seriously

Much like AGI, there will never be a moment where The Singularity™️ happens. There will be a bunch of goalpost moving, but eventually it'll just get to the point where the arguments that it isn't happening are so weak that people will stop making them.