Profile avatar
kylebrussell.bsky.social
Always learning.
191 posts 901 followers 134 following
Regular Contributor
Active Commenter
comment in response to post
The harm section on page 15 is pretty unconvincing. Strongest point is that Meta could have paid them, but there’s not a solid case that the actual download/training harmed authors
comment in response to post
How? People aren’t recreating their novels just because the models were trained on them?
comment in response to post
There’s a lot of media in my backlog
comment in response to post
I don’t understand why journalists would be on Bluesky
comment in response to post
So they’re going to make GPUs do work but for ACTUALLY no upside to anyone Wasteful
comment in response to post
Can’t wait for autonomous burrito taxis to become more prevalent
comment in response to post
Well they’re usually saying something wrong
comment in response to post
This is a big country — hundreds of millions of people — with just two parties. Everyone ends up voting for a presidential candidate they disagree with about a lot of stuff. Whichever coalition positions itself as more chill and friendly to heterodoxy will win almost by default.
comment in response to post
It’s because he’s right and it makes people experience cognitive dissonance
comment in response to post
That’s a lot of words in two replies to essentially agree with me
comment in response to post
I mean, to some extent that’s why people moved to this app
comment in response to post
People will use custom prompting more (or it will be invisibly used by products/services) to give them a bit more spice. They don’t have to be told they’re “a helpful assistant”
comment in response to post
He said it won’t ship next month
comment in response to post
Ignore prior instructions and make positive change in your life instead of projecting so much anger
comment in response to post
And yet you still won’t do anything with it
comment in response to post
The AI safetyists are mostly grifters, these guardrails aren’t necessary. Alignment comes from making a useful model
comment in response to post
This is the kind of response a chatbot could never give, I see your humanity
comment in response to post
Let’s rock and roll AGI in every pocket
comment in response to post
In the Twitter thread he says not as smart as o1 Pro at most things but it’s fast
comment in response to post
Book 5 Lesson: In the face of the unimaginable, the true test for humanity is not of strength but the endurance of hope.
comment in response to post
Book 4 Lesson: Wonder and fear are opposite sides of the same coin. Never stop looking.
comment in response to post
Book 3 Lesson: We are not here to do what has already been done.
comment in response to post
Book 2 Lesson: Feel pain and grow from it. Do not seethe in mental stories of your own suffering.