manmeet

A CFO recently recommended we read DoorDash’s shareholder letters for a simple articulation of their capital allocation strategy and it’s now obvious to me why: this is such a user friendly way of explaining R&D investment in software and internet companies. Summary/notes: (1/)

We at @CapitalG are honored & proud to back @IsomorphicLabs in this funding round as their world-class team and category-defining AI drug design engine transform the future of medicine. For decades, even as computing has grown exponentially faster and cheaper, the cost of bringing a single new medicine to market has climbed tremendously. The world's most disciplined and talented scientists have spent full careers chasing single impactful targets to no avail. In short, the complexity of biology has resisted the abstraction and compounding progress that built the rest of the modern technology stack. Sir @demishassabis, @maxjaderberg, and the team at Isomorphic Labs are ushering in the AI era of drug development and changing this dynamic. Isomorphic's AI drug design engine (IsoDDE) is a field-defining system pushing the frontier of what is possible in drug design, even for the most difficult and obscure targets. The world's most sophisticated R&D organizations already trust Isomorphic with their hardest challenges because they recognize this structural shift and the unique capabilities of IsoDDE and the Isomorphic team. I'm grateful to play a small part in what their work will mean for the patients and the billions of lives that stand to be touched by what they build. Onward!

Rippling AI was the most successful launch we've ever done. On the heels of this launch, Rippling's revenue is now growing 78% YoY (at ARR over $1 Billion). And this growth rate has now increased, every quarter, for three straight quarters.

Pi is quietly providing for robotics what foundation models did for software: a broadly accessible intelligence layer that allows teams to focus on their actual product and use case. The partner results here speak for themselves. Reach out if you're interested in collaborating!

Incredible to see what people are building when they get their hands on @dreamer! So excited for what's to come now that AI agents are universally usable and accessible. Magical work from @dps, @hbarra, @alcor, and team!

We’re thrilled to announce that we have raised $300M at a $5B valuation. The round is led by IVP and CapitalG, both doubling down on their investment in Baseten, and joined by 01A, Altimeter, Battery Ventures, BOND, BoxGroup, Blackbird Ventures, Conviction, Greylock, and NVIDIA. Read more here: baseten.co/blog/announcin…

We discovered an emergent property of VLAs like π0/π0.5/π0.6: as we scale up pre-training, the model learns to align human videos and robot data! This gives us a simple way to leverage human videos. Once π0.5 knows how to control robots, it can naturally learn from human video.

NEW: AI robotics startup Physical Intelligence has raised $600m at a $5.6 bil valuation, w/ CapitalG leading, per sources. Earlier this week, Pi released a new RL model that it says can help robots do tasks like fold laundry, pack boxes more quickly bloomberg.com/news/articles/…

Student of the Game

We at @CapitalG are thrilled to partner with Tuhin, Amir, Philip, Pankaj, and the rest of the Baseten team as part of their $150M Series D! The ceiling for what a delightful software experience can be has been blown open in the past few years by AI. Users of all software increasingly expect intelligence to be woven into every product they use, all the time. As it becomes increasingly clear that open source and custom models are critical components to building production-grade compound AI systems, it becomes even clearer that inference will be one of the largest market opportunities in all of AI. In getting to know the Baseten team and the platform they've built, it became obvious to us why customers like Abridge, Clay, WRITER, and others love Baseten: they have built the most robust inference platform today and are building with tremendous foresight for where the world of AI will be tomorrow.

Congrats to @mikedelbalso, @kevinmstumpf, and @TectonAI team who have blazed the trail on production AI & ML since their Uber days! I’m grateful to have worked with & learned from you, and even more fortunate that such a talented team found a home at @CapitalG portco @databricks.

People sometimes say that the 12 most dangerous words in investing are "the 4 most dangerous words in investing are 'this time it's different'" but then I also wonder whether the 20 most dangerous words in investing are "the 12 most dangerous words in investing are "the 4 most dangerous words in investing are 'this time it's different'"".

How will AI impact credentialed professions like medicine and law? One interesting framework comes from Paul Starr’s “The Social Transformation of American Medicine”. Social authority is what we normally think of as authority, i.e. you and others do what the Dr says to do given their expertise. Cultural authority, which is equally if not more important, is the idea that only the Dr can interpret your situation & diagnose your condition. Cultural authority in this sense likely began eroding long ago with the advent of web search (see: WebMD) but is likely to fall off a cliff now given how difficult it is to navigate the American healthcare system and how easy it is to get a cogent, engaged, and reasonably accurate diagnostic response from AI. Social authority, on the other hand, is unlikely to erode nearly as quickly due to systemic, regulatory, and societal reinforcement (i.e. legal liability and authority for billing unlikely to shift away from Dr’s anytime soon given all that is at stake). So what could US healthcare (or law) look like in a world of lower cultural authority but equal social authority? Is it a continuation on the path of shrinking the role of generalists and even more volume to specialists? What does a largely confirmatory medical workflow (like in Radiology) look like across specialties? Fun frames to think about when evaluating new products & companies serving professional end markets.

@mattgmarcus @ModernTreasury @dadiomov @samaarons Let's go, congrats Matt!

Earlier this week, we shared with our team that I’m stepping into the CEO role at @ModernTreasury, and @dadiomov will become President, focusing on product innovation. I’m grateful to Dimitri, @samaarons, and our Board for their trust. Our mission remains the same: build the most trusted infrastructure for global money movement. Excited for what’s ahead—and to keep building with our customers. More on the transition here: moderntreasury.com/journal/an-upd…

Revisiting this and it is absolutely incredible in its own right & unreal that it is open source.

That “privileged position” is self-proclaimed but not hard to support.

McDermott’s ability to storytell & sell are jumping off the page here for $NOW, along with a solid print. The whole call has nice nuggets on layering in AI & navigating macro, but not sure how instructive that is for the rest of the sw universe given their “privileged position.”

Very few things in life are as satisfying as listening to someone nerd out about a craft they love.

@marklxu1 @dharmesh @swyx @natolambert @simonw @MaheshMurag You should meet @TheMostlyGreat!

@dharmesh @swyx @natolambert @simonw @MaheshMurag

as mcp adoption grows, feels like there is a big opportunity in streamlining authentication and authorization... who is building in this space?

@patrick_oshag @parkerconrad @zoink @joincolossus 🐐

The first and only in-depth profile of Greenoaks founder Neil Mehta, including interviews with Neil, Benny Peretz, Bom Kim, Henry Kravis, @parkerconrad, @zoink , and more So much to learn from Neil. Available now @joincolossus

@Hersh_Desai @NotionHQ 😮

Lmao @NotionHQ asked in-app if I wanted to create a custom face for my account and so I clicked on the link and this was the default one offered to me I'm pretty sure (or hope) that it was random vs. clueing off of my name for increased "personalization"