Fun podcast on kernels, origins of @togethercompute, and how we’re using AI to build inference in an RSI inference loop with Together Compiler.
Thanks so much for having me @gabyllorenzi and @BSchech!
Congrats to @AlexKsendzovsky and team for their incredible launch.
One joy of seed investing is the chance to look into the future, which will be built by founders who create magic.
No investment has tought me this more than @BioComputingCo
https://t.co/ZLIIGLZmcP
today, we introduce a little mischief to the world.
for four years, we’ve been brewing real brain cells on electrodes to understand how they compute. now, we’ve turned those discoveries into a video model anyone can use.
from 0.5¢ a second, mischief 1 is at least 5x faster than its base and the cheapest image-to-video model on @fal and @OpenRouter. soon to be on @awscloud marketplace.
it’s not a delusion anymore. biocompute is real, commercial and this is just the beginning.
happy halloween.
Most of the attention in AI goes to the chips. Barun Kar and @rkhemani working on how those chips get connected, and they’re betting that network should be open instead of locked to one vendor. New episode of The Compute 100 with the @upscale_ai_inc founders.
Today’s guests are Barun Kar, CEO, and @rkhemani, Executive Chairman at @upscale_ai_inc.
Barun and Rajiv’s bet is that the AI era needs its own network: open, standards-based and built to connect AI chips to each other.
From switch silicon to supply chains, we discuss:
- How scale-up and scale-out networks work
- Where copper, optics and optical switches fit
- Why open standards matter for AI chips
- What a network operating system does
- How Upscale secures its supply chain
- Building culture while scaling to 300 people
Timestamps:
0:00 Introduction
1:05 Why AI needs a new network
5:21 Why cloud-era chips fall short for AI
9:05 Scale-up and scale-out inside an AI rack
16:09 The Auradine spin-out and Upscale’s founding story
20:52 Memory semantics versus packet switching
27:23 From Cavium and Palo Alto Networks to Upscale
32:23 Competition and the system-level advantage
35:55 What a network operating system does
37:01 Optics and optical switching
41:45 Selling to hyperscalers, new clouds and enterprises
47:40 How big AI networking could get
50:43 Securing supply across the chain
54:29 Building at the forefront of AI
57:40 Beyond networking: what comes next
1:03:38 Building culture in a fast-growing company
Listen to the show, and follow along for new episodes every Tuesday.
@CallosumAI raised a $100M seed on the idea that the future of compute is many different chips, each handling the part of the problem it’s best at. Great conversation with @achterbrain about how they’re building it. New episode of The Compute 100.
Today’s guest is @achterbrain , co-founder and CTO of @CallosumAI .
Jascha’s bet is that the future of compute is heterogeneous: different chips working together on different parts of a problem.
Beyond the multichip future, we discuss:
- What neuroscience taught Jascha about designing computing systems
- How chipmakers discover what their hardware is best at
- Where thermodynamic, biological, and quantum computing could fit
- Hiring people who welcome having their ideas challenged
- Building an AI infrastructure company in London
Timestamps:
0:00 Introduction
2:55 What heterogeneous computing actually means
11:02 From PhD research to Callosum
18:36 Turning heterogeneous compute into a product
23:24 Sourcing compute and working with chipmakers
28:53 Matching workloads to the right hardware
34:24 What heterogeneous computing can do today
38:29 The next waves of chips
48:02 Building the team
51:33 Why build in London?
56:11 Helping enterprises upgrade their compute stacks
1:01:04 Fundraising and building momentum
1:03:40 Where specialization wins, and where it doesn’t
Listen to the show, and follow along for new episodes every Tuesday.
i want to find an amazing electrical engineering professor to teach a room of the compute curious, the people who want to pivot into compute, and maybe some VCs about semis 101 - basic language, physics, systems, etc.
who should i have teach the class? and who wants to come!
Neurotech is having a moment, and SF is right at the center of it.
Big shoutout to @naveen101 for bringing the community together at #SFTechWeek!
Meanwhile, @AlexKsendzovsky holding court on the future of biological computing.
Not many people writing about chips actually trained as semiconductor physicists. @austinsemis did — grad work at Illinois, years at Intel — and it shows in every issue of Chipstrat. New episode of The Compute 100.
Today's guest is @austinsemis, the writer behind @chipstrat.
Austin started his career as an engineer at Intel before moving into startups and independent analysis. Today, he writes about semiconductors, optics, networking, and physical AI, bringing an engineer's perspective to what's happening across the compute stack.
We're thrilled to have him join us for a conversation on the state of the chip industry. Give it a listen!
Inference has become the dominant workload for AI. Our core thesis is that the future of inference will look very, very different from how it looks today.
No founder has expanded our thinking on what that might mean more than @AlexKsendzovsky.
Congrats on the launch to the whole team. Honored to be along for the ride into the future.
@NaveenGRao thinks power is the binding constraint on AI, and that the answer is silicon that runs neural networks closer to the way a brain does. Unconventional AI raised $475M at a $4.5B valuation to go find out. New episode of The Compute 100.
Today’s guest is @NaveenGRao, CEO and founder of @unconvAI. Naveen has been early to some of the biggest, most influential markets in technology, building Nervana and MosaicML ahead of his current venture: Unconventional AI. To Naveen, the future of compute means that systems will look much more similar to how the real world computes, as opposed to a paradigm we’ve all accepted as the best fit. Listen to the episode for an early look into what Unconventional AI is building and Naveen’s perspective on their journey ahead.
Did you know you can learn interesting things from computations in real rat neurons? I love what @AlexKsendzovsky and team are doing at @BioComputingCo are doing.
We encoded the @Twitch logo into a network of living neurons today.
Then we watched them respond.
This is where our AI software starts: learning how real neurons compute.
We invited @awscloud in for a tour. Check it out here:
https://t.co/6GtYijLrIX
@PatrickTBowen sees an opening the rest of the industry has missed: prefill, not decode, as the place optical compute wins first. That's the bet behind @neurophos . New episode of The Compute 100.
Today's episode is with @PatrickTBowen, CEO of @neurophos. We're thrilled to have him join us for a discussion about his journey from metamaterials to optical computing. Give it a listen!
The fastest company to ship production AI chips from scratch. @kleinerperkins is thrilled to back Gavin, Rob, Chris and the entire @etched team. We couldn't be more excited!
https://t.co/1kYqZA3jM1
I recently spent several days onsite at Etched. It was the most exciting and energizing business environment I’ve ever been in.
I left inspired, and wondering: “have I even begun to explore my personal potential?” People at Etched are exploring theirs.
They have open roles across the org. I believe deeply it’s a place you can do your life’s work.
One of their mottos is “Production is the product.” The team is off the charts talented, but they need even more great people to fulfill the mission.
As they said, “The racks won’t build themselves.” You should join them.
Racks on racks on racks! This team 🛳️🛳️🛳️
Congrats again to @robertwachen@UbertiGavin and the entire @Etched team. And huge welcome to all our friends for joining us along this insane journey.
And it's still day 1!
Three years ago, almost everyone said Etched couldn’t be built.
Today, they’re shipping production systems to Jane Street.
Huge congrats to the entire @Etched team, @UbertiGavin and @robertwachen for your stunning leadership. Thank you @sonyatweetybird for seeing from the outside what had become obvious on the inside, @stripesco_ for betting on the company at such a crucial moment, and @patrick_oshag for your steadfast championing of the mission.
A personal reflection on this milestone:
https://t.co/mWy6y4p5tF
Three years ago, everyone told us this was impossible when @PrimaryVC led the seed round.
Today, @Etched is shipping production racks to Jane Street.
Life-changing and extraordinary to watch this team turn a contrarian idea into working silicon, real systems, and real customer deployments.
Trust me on this one: they’re just getting started.
We've raised $700M at a $21B valuation from Jane Street, Kleiner Perkins, Sequoia, A16Z, Peter Thiel, BCV, and Blackstone.
We're also excited to share that we've shipped our first rack to Jane Street.