gaurav

notes to the void. some echo back.

[gaurav @gauravcy]

your ChatGPT could never come up with a sentence like this

[gaurav @gauravcy]

you don’t “deserve” anything, no one does. it’s just about what you can get in the end

[:p @UnkleDell]

spanish niggas be like intelligence artificial

[gaurav @gauravcy]

as opposed to individuals, governments should be considered guilty until proven innocent

[gaurav @gauravcy]

time passes anyway, do the thing.

[gaurav @gauravcy]

most of modern day stoic advice on internet is just advice on “how to cope”

[gaurav @gauravcy]

attention is all you need and attention is all you have

[gaurav @gauravcy]

pain is real, meaning isn’t

[Dwarkesh Patel @dwarkesh_sp]

Did a very different format with @reinerpope – a blackboard lecture where he walks through how frontier LLMs are trained and served.

It's shocking how much you can deduce about what the labs are doing from a handful of equations, public API prices, and some chalk.

It’s a bit technical, but I encourage you to hang in there - it’s really worth it.

There are less than a handful of people who understand the full stack of AI, from chip design to model architecture, as well as Reiner. It was a real delight to learn from him.

Recommend watching this one on YouTube so you can see the chalkboard.

0:00:00 – How batch size affects token cost and speed
0:31:59 – How MoE models are laid out across GPU racks
0:47:02 – How pipeline parallelism spreads model layers across racks
1:03:27 – Why Ilya said, “As we now know, pipelining is not wise.”
1:18:49 – Because of RL, models may be 100x over-trained beyond Chinchilla-optimal
1:32:52 – Deducing long context memory costs from API pricing
2:03:52 – Convergent evolution between neural nets and cryptography

[gaurav @gauravcy]

Couragemaxxxing

[gaurav @gauravcy]

You don’t build totalitarian version of Palantir. It is the final state of every “Palantir” to become totalitarian.

[Bojan Tunguz @tunguz]

Startup idea: a non-totalitarian version of Palantir.

Who’s building this?

[gaurav @gauravcy]

@DavidDeutschOxf @webdevMason What are the other good ways of criticising what you want now, apart from long term?

[David Deutsch @DavidDeutschOxf]

@webdevMason Choosing a destination as a means of setting shorter term priorities is a bad idea. It makes it hard to discover your mistake until you're dead.

Hash out what you want NOW—much rarer & harder than it sounds—and think about the long term as one of many ways of criticising that.

[gaurav @gauravcy]

It is what it is

[mert @mert]

books are crazy because theyre like twitter threads but by non-retarded people

[gaurav @gauravcy]

every day, the most courageous thing you can do is be in a fucking good mood

[gaurav @gauravcy]

It’s not a good idea to remember everything perfectly

[Gabriel Kreiman @gkreiman]

Imagine a future where you can REMEMBER EVERYTHING.
Every email, every person, every conversation.
Introducing Engramme.
Our vision is to endow humans with perfect and infinite memory.
All your memories come to you. No more searching or prompting.
https://t.co/LX22KMG2lH 🧠

[gaurav @gauravcy]

never reject the muse

[gaurav @gauravcy]

That’s the only right way of bird watching

[michael @FilledwithUrine]

bird watching while not knowing any bird names thinking oh yeah that’s gotta be a rare one