Cameron retweeted
What is jhana?
Announcing our new paper in Mindfulness (@springernature):
“Toward a Unified Account of Advanced Concentrative Absorption Meditation. A Systematic Definition and Classification of Jhāna”
We present the most comprehensive review of advanced concentrative ...
🧵
Cameron retweeted
here's a 10-minute attempt to summarize my 300 page book @introspectVV into ~500 words. didn't precisely meet the brief of "in one breath" but i'll try again multiple times
Cameron retweeted
post-rationalism doesn't require woo, but one of the nice things about noticing that scientific materialism is unfalsifiable and bootstrapped out of the void is that you don't have to accept it's view of a disenchanted, mechanistic world as unassailable and self evident
Cameron retweeted
"Alignment" isn't a thing. It's a philosophical fantasy of a general first-principles guarantee of safety. That will never happen.
You get safe devices with engineering, based on science.
Current AI practice is neither engineering nor science. betterwithout.ai/fight-unsaf…
Cameron retweeted
Utilitarianism is codified psychopathy, and its application to altruistic endeavors an oxymoron.
It is therefore no surprise that EA consistently leads left-lateralized types who imagine themselves saintly down a Luciferian path.
Cameron retweeted
“AI Alignment” is a red herring.
The systems—as such—aren’t the concern.
They’re not close to “conscious” agency.
The present x-risk lies primarily in accelerating the externalities entailed by our capacity to centrally attract, manipulate, and shape human attention at scale, as well as the immense amplification of leverage this grants to long tail actors with nothing to lose.
Cameron retweeted
As a functional distillate over the entire expression-space of human knowledge, this is *not* surprising.
It's useful, but to the extent it displaces the processes that produced the artifacts upon which it rests, it will starve itself of input sources beyond its own exhaust.
unavailable
Cameron retweeted
🆕🤖 How in 2014 I figured out that "neural network" image classifiers weren't doing what they seemed to be; how they were cheating instead; how that was confirmed by others' independent experiments; and what this may imply for GPT-like systems. [Link in next tweet ->]
Cameron retweeted
The secret to happiness turns out to be not doing a specific fast clenching mechanic called tanha that we’ve known about for 1000y+, is easily reproducible by tons of people, and we don’t teach people about ever. It all feels like a cheesy plot point from a fiction novel
Cameron retweeted
These are the most detailed images of a human cell to date, obtained by radiography, nuclear magnetic resonance and cryoelectron microscopy.
Cameron retweeted
When human minds are miswired into becoming nerds or narcissists, they may not notice how their problems are rooted in themselves and go on a quest to debug the universe instead. The nerd questers are called analytic philosophers, narcissist questers are continental philosophers.
Cameron retweeted
So, for my first tweet, I choreographed this thing with @BostonDynamics. Thanks to the team for their incredible work.
Our whole crew got together to celebrate the start of what we hope will be a happier year. Happy New Year from all of us at Boston Dynamics. youtu.be/fn3KWM1kuAw
