Sabitlenmiş Tweet
sisyphus bar and grill
38.1K posts
sisyphus bar and grill
@itunpredictable
blog poaster @amplifypartners @readtechnically
Corinth Katılım Temmuz 2012
715 Takip Edilen19K Takipçiler
Can't think of a more Amplify thing than this, especially on a day where the megafunds are out bickering. @sarahcat21 is one of a kind
Tim Rocktäschel@_rockt
So excited to have Sarah support us at @Recursive_SI. She has been following our founder's Open-Endedness research for many years... even before it was cool. I have rarely seen a VC that is that deeply knowledgeable and opinionated (and correct) about future AI trends.
English
@itunpredictable tried it once, actually better than paper I promise - had to abandon my paper copy on a road trip
English
Here’s an idea. Why don’t you give me half the money you were gonna invest, we’ll go out back, I’ll kick you in the nuts, and we’ll call it a day.
Sheel Mohnot@pitdesi
In case you were wondering where we're at in the cycle, I just heard about a seed round where the founder is selling secondary
English
It's 2026, you're on a JetBlue plane called "Take a Menta Health Day"
English
Today I’m quite grateful for the VC moms WhatsApp group & the support I’ve received from other women balancing their commitment to work & motherhood.
Turns out there is a Group Chat among top tier VCs. We just discuss potty training & kid friendly hotels…
Happy Mother’s Day!
English
English
walking around the office with a big Eye of Sauron taped to my laptop
bullet chess clocks in conference rooms
energy for next N quarters
we are on the exponential, if you are of sound mind and body u must be locked in
dont blink
English
@yrechtman I bet your wife doesn't have these slop cannons
English
@skooookum @hopes_revenge cap. nobody has ever done 60 on the bqe.
English
English
@MaxDiCapua and I left @Uber late last year to build @19thLabs.
You give it data and a task. You get a small, fast specialist model that beats frontier LLMs on that task.
English
Big if true. Is this legit?
Alexander Whedon@alex_whedon
Introducing SubQ - a major breakthrough in LLM intelligence. It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA), And the first frontier model with a 12 million token context window which is: - 52x faster than FlashAttention at 1MM tokens - Less than 5% the cost of Opus Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention). Only a small fraction actually matter. @subquadratic finds and focuses only on the ones that do. That's nearly 1,000x less compute and a new way for LLMs to scale.
English
@nikillinit au contraire. it's embarassing to know.
youtube.com/MorePlatesMore…
English
@itunpredictable maybe this is embarrassing to ask but what is MPMD
English
fine, let's talk about peptides
Peptides are creating a parallel system to healthcare where it's pretty hard to tell if anything is working and even harder to figure out who should be looking out for issues
Every part of the value chain is incentivized to tell you that your problems will be solved
But it's worth understanding why this phenomenon is happening - which is why this newsletter is a call for discussion. I'd love for people to tell me their thoughts about peptides so we can have a nuanced discussion
outofpocket.health/p/fine-lets-ta…
English