Sarah Guo

America wins by recruiting builders and innovators: Einstein, Grove, Brin, Huang. Immigrants start companies, create millions of jobs, and strengthen our global leadership—benefiting generations of native-born Americans. Immigration done right is America First. 🇺🇸

vibes from @conviction’s A+ talent team today https://t.co/ivIZgM8P6p

What am I missing? https://t.co/zfypb2p3ge

@SaidAitmbarek no this was it implementing something that used my local fs but not shipping a fs, so it literally worked…on my machine

A coding agent essentially saying “it works on my machine” is just the funniest

memo writing vibe https://t.co/RErBb9PYOw

@t_xu wow, congratulations!

“The most talent-dense set of teams I’ve ever seen at a demo day” 💅

plane suggestions? books, albums, papers welcome here is what I have https://t.co/IqwYb9t4ZT

.@conviction Embed demo day T-30 minutes https://t.co/eJZJq8uoBt

the one in which me, @mims and @timkhiggins talk about tech’s midlife crisis and how them young’uns use AI

VC’s do you prefer company

We need more data for ai x science. @conviction is backing novel data generation ideas (at a scale that supports training big models) Somite’s “capsules” (via Allon Klein, prof systems bio at Harvard) is one such innovation — 1,000x’ing stem cell differentiation data https://t.co/DOHS9ZH1SM from @Forbes https://t.co/FVjCgU4xcW

my partner @pranavreddy makes fun of me for committing to main from a github mobile client like a savage but what are code agent interfaces but UX for this?? I live in the future also I’m sorry it’s your fault you gave me the power

.@NoPriorsPod with @LuisvonAhn founder of @duolingo Topics: *gaming as the future of education *motivation as the hard problem of learning *going ai-native *unhinged brand *duolingo for chess, math and more languages https://t.co/Ov4qmsQ1fj On YouTube and audio platforms: https://t.co/wWSmJNbHBK

.@fidjissimo is an amazing leader and this is a huge get

this @NoPriorsPod was so fun: Josh Goldman of @KoBold_Metals about rare earth minerals, lithium, taking AI to real world industries like mining, and the role of a human scientist/expert in the age of AI https://t.co/G0Mwgw5jNL Full interview on podcast platforms and YouTube: https://t.co/62ocs3za6D

as someone who mostly learned to write bad code in a real world environment with a senior engineer growling at me…and later tried to go back and take a couple “foundational” CS classes at both Penn and Stanford learning new things with Cursor, Devin, Claude feels 💯 better or who knows if I’m learning now! “If it works it’s correct”

1/ useful to catch up with friends and judge the Launchpad competition @ RSAC2025. There was a clear shift from CISO skepticism on AI to embracing it as mission-critical weapon. Some takeaways (and questions) in thread 👇 https://t.co/4sPryZgRct 2/ AI is now both lock-pick and shield. Attackers can cheaply craft polymorphic phishing & deepfakes; defenders must counter with real-time AI detection or be overwhelmed by volume. The arms race is underway. Cyber is one of the first and most real abuses of LLMs 3/ SOCs are going autonomous. AI triages alert floods, stitches evidence, investigates more deeply (a combo of log parsing and securit…

1/ Possibly biased but @gradypb and @sonyatweetybird dropped some gems at @Sequoia’s AI Ascent today: https://t.co/trxpOu4pKz 2/ TAMs for software and services are both wide open and much bigger than a decade ago—companies are shifting from “copilots” to “autopilots,” from selling tools to selling outcomes. 3/ Tech distribution physics are wild now. Salesforce started when 200M humans had internet; now your grandma is streaming TikToks from her rural farm—5.6B humans strong. 4/ The race for markets traditionally have been “incumbents have distribution, startups have tech.” But given these new physics, Pat thinks in the AI age it’s more “fo…

a few of my favorite tidbits from the newest @NoPriorsPod with @OpenAI o3 team members: “Humans are a super expensive tool call. If you’re a model, you can either ask the meat bag over here for help—or you could’ve browsed and read a hundred papers instead.” -@mckbrando “If you want the model to be good at collaborating with people, maybe a not-too-bad starting point is making it good at collaborating with other models. Someone should do that.” -@mckbrando “How long should you wait? Ideally, as long as it takes for the model to give you the correct answer—and models should understand their own uncertainty enough to decide.” -@ericmitchella…

quick trip to Hill and Valley. Thanks to @jacobhelberg on building a needed, ever-stronger bridge between the policymakers and builders in tech couple off script notes below https://t.co/2mkm258fXy 1/ to start: I’m impressed by Karp’s composed handling of anti-Israel, anti-AI protestors — real belief in free speech as an American right. capitol police far too slow 2/ to believe in the power of AI is to believe the west must lead in AI. innovation (and meritocracy) is the root of America’s greatness, and we must re-enable that 3/ America’s strategic bottleneck in AI isn’t chips—it’s electricity. without abundant, reliable power, we can’t r…

@thesephist *computer

ai is uniquely something you just use more and more and more of

really explains a lot of interactions with @mvernal https://t.co/ZfKDr8OCNo

Here’s a sorting hat for who’s on your side: who is willing to have an uncomfortable conversation for you? It’s easy (and cheap) to cheerlead

News: ChatGPT's Deep Research for free! Courtesy of lighter weight models :) 🆕 @NoPriorsPod with @OpenAI's @isafulf on: *how they built it (and improve it) *what data we need to train better agents *agents going from read --> write, safety *how she uses Deep Research herself https://t.co/K1EBqhdHmh Full interview on YouTube or wherever you get your podcasts https://t.co/Aws6t3EvkW https://t.co/kp7zmttUXE

@jasoncitron @discord congratulations on building an amazing product (and business!)

simple tests of organizational health: - how much performative “work”? - are people genuinely excited? - do people trust one another?

ok I’m not exactly a Luddite, but I do want to be able to occasionally think “alone” https://t.co/R6DbvJcLWe