←
Will Reed2,132 posts
Will Reed avatar

Will Reed

@willreed

gp @sparkcapital

◷ Joined October 2015▧ 223 Photos and videos
Will Reed avatar
Will Reed@willreed·Jun 23

Baseten will be one of the defining infrastructure companies of this era. Its Inference Cloud powers today's most advanced AI apps, and as post-training becomes inseparable from inference, they're working with app builders to create closed-loop learning systems that turn production data into compounding model improvement. We're proud to have led or co-led its $40M Series B, $75M Series C, and now its $1.5B Series F. It's Spark's second largest investment ever, behind only Anthropic. The team is special and the market still underestimates the size of the opportunity. We couldn't be more excited for what's to come.

Tuhin Srivastava @tuhinone Jun 22 Announcing our Series F Today, we are thrilled to announce Baseten’s $1.5B Series F, led by Altimeter Capital, Conviction Partners, and Spark Capital, co-led by Sands Capital and Wellington Management, with participation

Dannie Herzberg avatar
⟳ Will Reed retweeted
Dannie Herzberg@DannieHerz·14h

Today we’re bringing Baseten inference to Microsoft Foundry. With Baseten on Foundry, enterprises get our inference performance alongside the security, governance, and billing they already rely on in Azure. Thank you to the Microsoft Foundry team for building this with us. baseten.co/blog/baseten-is-n…

Mario Gabriele avatar
⟳ Will Reed retweeted
Mario Gabriele@mariogabriele·Nov 17

.@tuhinone spent three years building a platform for data scientists, until a customer arrived with a fine-tuned Stable Diffusion model that took off on Hacker News.  Within six weeks, he and his co-founders had thrown that work away to focus on a single piece: serving models. That piece became @Baseten, which now runs inference for companies like Cursor, Abridge, and OpenEvidence across roughly 20 clouds. In June, a $1.5 billion round valued it at $13 billion, though its revenue forecasts have been off every year since 2023, in 2025 by 4X. In this conversation, Tuhin explains why he believes inference is the COGS of AI, why companies may end up running dozens of specialized models, and why, even in a world with AGI, “the only thing left to do will be inference.” (02:06) What Baseten is building and how it began (04:47) Finding balance in a fast-growing startup (06:45) Why inference is the “final market” (08:34) Tuhin’s path to founding Baseten (12:06) What Australia gets right (13:43) The opportunity Baseten found in inference (20:34) How customers use Baseten (29:00) Managing compute across clouds (30:40) The compute crunch (32:11) Financing inference at scale (35:40) Why inference is hard to build in-house (36:41) Who Baseten wants to hire (39:46) Hard work and skill vs. luck (42:20) Final meditations

44:17
Shiv Rao, MD avatar
⟳ Will Reed retweeted
Shiv Rao, MD@ShivdevRao·Nov 16

When I’m on call, most of my work is a chart biopsy. What if a fellow was available around the clock to review the chart, present the patient, and help coordinate care? At The Conversation last month, we showed what that looks like: the @AbridgeHQ Assistant turning my goals into a plan, working across systems, and coming back when it needs my judgment. For me, the purpose is presence. Full attention on the patient, and confidence that the plan we make together gets carried through. We’ve talked about the Learning Health System for twenty years. We’re building it now with our partners. Full demo: abridge.com/blog/learning-he…

0:43
Baseten avatar
⟳ Will Reed retweeted
Baseten@baseten·Nov 16

Introducing the Baseten Voice Agent Stack: optimized infrastructure to run and customize voice agents. Every model and the orchestrator are hosted for low end-to-end latency, each component scales on its own, and you can swap or post-train any model. We built Baseten Realtime Voice on it using open-weight models: 73.5% on τ²-Bench Voice, more than double the cascaded baseline, with sub-second turns at ~$0.01/min. Both are in private preview now: baseten.co/blog/introducing-…

Post attachment
Phil Howes avatar
⟳ Will Reed retweeted
Phil Howes@saltyph·Nov 13

we built an inference engine for deepseek v4.1-flash from scratch. the core engine is just rust + cuda. no pytorch, no sglang, no dsls, etc. ~1740 TPS decode on 4xB300, 2060TPS on 8xB300. 1M token prefill in ~12s. it's up on github

Ashlee Vance avatar
⟳ Will Reed retweeted
Ashlee Vance@ashleevance·Nov 13

About a decade ago, a guy named Jim Belosic started a hobby project to merge a Honda Accord with a Tesla. The result was something called Teslonda. It was very fast. Possibly too fast. As he was building, Jim needed access to some decent metal shaping tools. He eventually decided to buy a $750,000 laser cutter and then he figured there were more people like him that also wanted access to good tools. So he started renting his laser cutter to others. All of this led to the formation of a company called SendCutSend. It's now making about 45 million parts per year for 300,000 customers. In many ways, it has become the US's answer to China when it comes to cutting, shaping and finishing metal parts. This is the deepest look at the SendCutSend factory and Jim's backstory that has ever been done. We went to his childhood home, met mom and dad (they're lovely) and drank Old Fashioneds at a saloon because that is the type of committment I am willing to put into being a video show host. Come watch "China, Meet Jim" - youtube.com/watch?v=sC4JvyhB…

Ben Liu avatar
⟳ Will Reed retweeted
Ben Liu@BenjamineYLiu·Nov 13

Thanks @bradloncar and @BiotechTV for having me on to share more about @formationbio

𝐅𝐫𝐨𝐦 𝐍𝐘𝐒𝐄: @formationbio announced a $600M fund it will use to acquire 10–15 drug programs over the next 3–5 years - a step in its plan to make drug development more efficient and become 'the next pharma'. @BenjamineYLiu Watch the full video: biotechtv.com/post/formation…

5:01
Scott Kazmierowicz avatar
⟳ Will Reed retweeted
Scott Kazmierowicz@scottwkaz·Nov 12

Today we’re launching Cardless AI. Cards that talk. Cards your AI agent can spend with. And Parallax, our model for catching the AI-generated identities most fraud systems miss. cardless.com/blog/meet-cardl…

Jason Carman avatar
⟳ Will Reed retweeted
Jason Carman@jasonjoyride·Nov 12

Zeromatter built a simulation platform from the ground up: their own renderer and physics engine, simulating cameras, LiDAR, radar, and thermal in real time to train autonomous systems. Today they're out of stealth. Honored to work with Ian & the team.

4 years ago, Zeromatter started with a single bet: building autonomous machines would be faster, safer, and better done first in a virtual world. Today, we’re coming out of stealth to show what that bet has become.

1:40
Zeromatter avatar
⟳ Will Reed retweeted
Zeromatter@ZeromatterTech·Nov 12

4 years ago, Zeromatter started with a single bet: building autonomous machines would be faster, safer, and better done first in a virtual world. Today, we’re coming out of stealth to show what that bet has become.

1:40
Shiv Rao, MD avatar
⟳ Will Reed retweeted
Shiv Rao, MD@ShivdevRao·Nov 12

Health systems have put years of clinical expertise into defining how they can best care for their patients, and those guidelines should be accessible when a clinician is making a decision. Today, @AbridgeHQ announced support for local pathways, a new capability within the Abridge Assistant that brings a health system’s preferred care plans into responses, alongside relevant medical literature and contextualized to the patient. With local pathways, a clinician doesn’t have to go searching for the next step. We are bringing that institutional knowledge into the same Assistant, and into the same workflow, that clinicians use to understand the patient and consult the evidence. Read more here: abridge.com/press-release/lo…

Post attachment
Shiv Rao, MD avatar
⟳ Will Reed retweeted
Shiv Rao, MD@ShivdevRao·Nov 10

At @AbridgeHQ, it is a privilege to serve Veterans and the clinicians who care for them.

2:03