Sarah Guo

next steps for the big labs feel analogous to being oil barons after the easy wells have been tapped. early capability gushers made fortunes (won markets like consumer asst, code), now the remaining oil requires deep-sea rigs or fracking: still possible, but a slow/costly grind

big opportunity for aggressive tech M&A. the AI step change in what you can build catching many companies flat footed (see: "death of SaaS" meme) as they lack the talent/organizational will to adjust quickly...but customer value creation is large enough to justify acquisitions

Healthcare AI is the most underrated story: adoption is outpacing other “early adopter” industries, it leapfrogs systems that couldn’t realistically be upgraded, and it’s not about squeezing costs — it’s tackling the shortage of people that’s already blocking access to care also: scribing is just one slice there’s taking notes…and then there’s providing care: making good decisions, communicating with patients, navigating admin to get treatment/meds modern medical care is sometimes already great, but it is not evenly distributed

As you start to walk on the way, the way appears

open source is the best. so many video editors and yet ffmpeg remains unrivaled

who should we interview about nuclear on @NoPriorsPod?

some of my favorite people for the early stage “haven’t scaled well.” eg first sales folks, recruiters. they’re worth their weight in gold and are gluttons for punishment doing repeat early, high agency problem solving work. in the age of AI, maybe they get to do that at scale why does it usually break? they’re doers, but they don’t create systems to manage lots of people, or they don’t manage the communication complexity, and imperfect and indirect incentives of big companies well venture firms like to talk about hiring “stage specific leadership” I think this will evolve. seeing lots of SaaS execs struggle within the super high growth an…

what to do with 12 hours in Rio?

Big companies tend to reward the best politicians, not performers.

my longtime friend, AI legend @AndrewYNg joined me and @eladgil on @NoPriorsPod this week -- he thinks AI can really help with investing, but it can't get at some of the people judgement or transparency borne of relationships (yet) https://t.co/iaYwkxLR5w full interview anywhere you get your podcasts, and on youtube https://t.co/o9gTmefcf3

The era of superintelligence is here. Didn't predict the medical field would be first. Amazing work by team @openevidence in scoring a perfect 100% on the US Medical Licensing Exam https://t.co/9J8vOYGzHy the average human doctor scores in the 70s without AI assistance. care is about to get a lot better https://t.co/HI4tFgCBJK

https://t.co/ba5MdfbDsL

.@conviction team giving @eladgil a run for his money with darts vibe check on investment https://t.co/whnejUEvcV

Congratulations @SolaAI_ on their launch and $21M of funding! Everyone’s talking about enterprise AI agents and computer use, but no one can tell you how this actually gets into real, economically valuable workflows. The answer is to make them self-serve and "teachable," so end users can rapidly train automations for everything they do without touching a line of code. Sola is the first platform where an operator can deploy on day one, scale to tens of thousands of runs, and never wait on a developer, FDE, or “center of excellence.” Long term believers in founders @jess123555 and @NeilDeshmukh. They're hiring! happy to partner with @kimber…

it’s a pretty heady time in AI, but fundraising is not core business progress. It can serve as a constant distraction — do not let it be only validation that counts in the end- customer love, fat margin, then ring that bell in the public pen

0/ as a formerly rebellious kid who valued control with a…highly independent child, I was thinking about agency and AI today. bear with me 1/ We value control even when it changes nothing. Mauricio Delgado’s “Value of Control” experiments: people guessed if a hidden number was >5 or let a computer guess. Odds were identical (~50%), yet most paid a cost to guess themselves. 2/ In fMRI scans, anticipating the chance to choose activated reward circuits—ventral striatum, medial prefrontal cortex—even when choice didn’t improve outcomes. Choice itself triggered a neural reward. We biologically want it! 3/3 If more people simply follow model…

0/ we should be answering so many curiosity questions with ai research right now. some I have below! If it’s been done pls tell me 1/ when do algos get upgraded ? Use AI-assisted code parsing to scan open-source history, detect algorithm implementations, and track when they’re replaced. Example: MD5 vanished from major projects as SHA-256 takes over, quicksort swapped out for introsort or timsort in language libraries. 2/ what code dialects exist? Use AI to extract and compare style and structure patterns across thousands of projects, grouped by type—scientific computing (NumPy, Bioconductor), games (Godot, libGDX), data pipelines (Airflow,…

These are some bummer charts for the 16-39yo’s (self included) Is there interesting research on how to increase these traits? https://t.co/L8cW5ShGiR

@_opencv_ @adhoc97 wouldn’t that be meta

1/ ONESHOTTED? In Snow Crash, a few words could change how someone thought forever. That’s fiction, but it feels close now? Spend enough time working with AI and your style changes. Your sentences and ideas compress into recognizeable patterns. This is being “oneshotted.” 2/ The bar for attention keeps rising, and AI raises it faster. Anyone can summarize now. The new muscle is being the *unpredictable, but clearly correct next token.* 3/ What emotionally resonates, but takes risk? What analogy from far-apart fields connects? What is the skeleton key to an industry — but runs counter to the narrative? Say something people didn’t forsee bu…

1/ we're extending the application deadline for @Conviction Embed, our grant program for AI-native startups to 8/14 it's the best time in history to start a company https://t.co/QrZfZ5SFTS 2/ for fun, we've added to our list of plausible schemes. first up, just Use The Recipe https://t.co/77h4Bon19Y 3/ Machine-Verified Trades https://t.co/07n7PoFe1z 4/ The OfficeVerse https://t.co/Me9PXkVczs 5/ Reasoning Infrastructure for Companies https://t.co/1z09FGt8n5 6/ Vertical Robots: Data In, Work Out https://t.co/y8djJlWdu4 7/7 Full list of startup ideas and application at https://t.co/gdxho7JVX2

ai, bio, robotics, energy these revolutions will drown out everything else over the next twenty years

the feeding frenzy for ownership in the AI labs has spawned a set of bottom feeding multi layered SPV brokers that have no relationship with the company, and straight up grifters. careful of that nonsense.

.@eastdakota, cofounder CEO of @Cloudflare on the newest @NoPriorsPod today about how content creators should be paid by ai labs for filling in the “swiss cheese block” of human knowledge, the end of search, advice for web publishers, and different futures for the internet https://t.co/YCLMciuUZi Full interview https://t.co/rbl00rJDlh

1/ models are getting better at code! how you use agents can also have 10x impact. who is using coding agents internally in "advanced mode"? dm for @conviction-hosted dinner invite on this topic some notes from conversations w/ @conviction's startup teams about best practices: 2/ front-load context. keep an https://t.co/4D2J2549pN in the repo with project purpose, architectural guidance, file structure, style guide, and test commands so agents write to your standards 3/ phase the work. plan → implement → review → refine --> commit. have the agent explain design choices and catch bad assumptions early 4/ build/use validation in the loo…

you can do more than you think AI opening up more people to understand that every day. Very exciting

https://t.co/gdxho7JVX2

.@chaidiscovery is *the* AI lab transforming end-to-end biological discovery into an engineering problem. investor from the beginning, and a huge believer in @joshim5 @jackdent @Mattmcpartlon1 @j_boitreaud and team - more flops for the fire

@dnlkwk https://t.co/PtSHs1Ujla

give me emdashes back