sarah guo // conviction
@saranormous
startup investor/helper, founder @conviction. accelerating AI adoption, interested in progress. tech podcast: @nopriorspod
Some personal news: We raised $100M for @w_conviction, an new early-stage venture capital firm purpose-built for Software 3.0 sarahguo.com/blog/conviction
Congrats to my friend @krisfredrickson! I wholeheartedly recommend him to founders Footnote: Kris was my first boss — who saved me from myself, taught me so much about finance, gstyle, humans and the grace of getting shit done while keeping your cool
Some news - I raised $175M for Verified Capital with a plan to back transformative technology companies at any stage of their journey. Thanks to @_IainMartin for the writeup in Forbes. forbes.com/sites/iainmart…
sailor moon x ai supremacy wasn’t on my bingo card for this year
newish fun board games that don’t take a huge amount of time to learn to play, and can be finished in 2H?
Congratulations @polynoamial and OpenAI’s math reasoning team on their IMO Gold 🏆
Today, we at @OpenAI achieved a milestone that many considered years away: gold medal-level performance on the 2025 IMO with a general reasoning LLM—under the same time limits as humans, without tools. As remarkable as that sounds, it’s even more significant than the headline 🧵
kimi2 median latency on OpenRouter: always fastest @basetenco
I'm at ICML for the week!! come find the @cartesia_ai booth to chat about architectures, tokenizers, voice AI, etc @sukjun_hwang and @fluorane will also be around to talk about H-Nets 🙌
"Raw intelligence is brain without hands," says Sarah Guo (Conviction VC). Large companies struggle with the velocity needed to deeply understand user needs and train AI for specific use cases. It's about AI's speed meeting real-world application. Watch the full episode from…
third time’s the charm! congrats windsurf/cognition
lol google sheets has AI before it has what-if analysis
fixing AI slop turns out to be as painful as fixing your own slop
few things as productive as making a 3-item to do list and just forcing your way through it on a flight
.@tbpn is faster than most traditional journos now and the whisper network is…not very much behind ours
We've been tirelessly working for the last few years to change how machines think. H-Net unlocks the next step of architecture scaling, bringing together many of the lessons learned in the last 5 years of SSM research to break fundamentally new ground in learning end to end…
Tokenization is just a special case of "chunking" - building low-level data into high-level abstractions - which is in turn fundamental to intelligence. Our new architecture, which enables hierarchical *dynamic chunking*, is not only tokenizer-free, but simply scales better.