←
Mike Schroepfer1,634 posts
Mike Schroepfer avatar

Mike Schroepfer

@schrep

founder @Gigascale, Sr Fellow/fmr CTO @Meta, founder @AdditionalVent. Let’s go build!

◷ Joined May 2007▧ 89 Photos and videos
Andrew Davison avatar
⟳ Mike Schroepfer retweeted
Andrew Davison@AjdDavison·Nov 12

This is a really good interview about Fractile and a clear vision of accelerating AI inference via more memory bandwidth. I met Walter at ICRA 2022 when he was presenting some of the final work in his PhD in robotics! The insights from Sir Robin Saxby who was the founding CEO of ARM are fascinating too... a very different British company from a very different time, but astoundingly successful in achieving his vision of becoming a global standard.

Walter Goodwin believes the world's most capable AI models are held back from delivering on their potential by the chips of today. Yesterday his company, Fractile, announced a $720M Series C to fix that. Walter, an Oxford PhD, founded Fractile in summer 2022, on a bet that there would be a huge dividend to speed in running AI models. The round is led by Thrive Capital, Founders Fund, Lightspeed and Redpoint, and comes six months after Fractile's $220M Series B. We sat down in London to discuss the future of semiconductors alongside Sir Robin Saxby, the founding CEO of ARM, who started one of the largest companies ever to come out of the UK with 12 engineers in a pub. Congratulations to @goodwin_ml and the @fractile_ai team. Enjoy this one! TIMESTAMPS 0:00 Intro 1:24 Starting Fractile 6:01 Inference Is the Profit Center 9:39 Chips Over My Dead Body: The Founding of ARM 18:45 The Art of the Opinionated Chip 25:09 Breaking the Memory Wall 32:15 How ARM Picked Winners 38:23 Robin Calls a Correction 42:51 The Stickiness of the Transformer 46:49 Is There a Limit to AI Demand? 56:11 How TSMC Became the Standard 57:44 AI-Designed Chips 1:03:52 Brutal Honesty

1:13:01
Finn Murphy avatar
⟳ Mike Schroepfer retweeted
Finn Murphy@FinnMurphy12·Nov 12

Walter Goodwin believes the world's most capable AI models are held back from delivering on their potential by the chips of today. Yesterday his company, Fractile, announced a $720M Series C to fix that. Walter, an Oxford PhD, founded Fractile in summer 2022, on a bet that there would be a huge dividend to speed in running AI models. The round is led by Thrive Capital, Founders Fund, Lightspeed and Redpoint, and comes six months after Fractile's $220M Series B. We sat down in London to discuss the future of semiconductors alongside Sir Robin Saxby, the founding CEO of ARM, who started one of the largest companies ever to come out of the UK with 12 engineers in a pub. Congratulations to @goodwin_ml and the @fractile_ai team. Enjoy this one! TIMESTAMPS 0:00 Intro 1:24 Starting Fractile 6:01 Inference Is the Profit Center 9:39 Chips Over My Dead Body: The Founding of ARM 18:45 The Art of the Opinionated Chip 25:09 Breaking the Memory Wall 32:15 How ARM Picked Winners 38:23 Robin Calls a Correction 42:51 The Stickiness of the Transformer 46:49 Is There a Limit to AI Demand? 56:11 How TSMC Became the Standard 57:44 AI-Designed Chips 1:03:52 Brutal Honesty

1:13:01
Drew Baglino avatar
⟳ Mike Schroepfer retweeted
Drew Baglino@baglino·Nov 12

@SheldonKimber thank you and the entire Intersect team for the close collaboration and commitment to bring a next generation utility scale solar inverter to market with @heronpower!

We couldn’t be more excited to announce that our first lighthouse partner, Intersect, has reserved 1.5 GW of Heron Link manufacturing capacity for solar. This has been many months in the making. Intersect worked closely with our technical and commercial teams to ensure Heron Link meets the needs of large-scale energy developers. Thank you, Intersect, for leaning in with Heron as we bring a step change in power conversion to our industry. We can't wait for the day we tour an operating project powered by Heron together. For more, see our release: heronpower.com/news-stories/…

Post attachment
Mike Schroepfer avatar
Mike Schroepfer@schrep·Nov 11

In a data center, chips are the biggest power draw. Inside one, it's moving data to/from memory. @fractile_ai stacks DRAM on top of logic. A direct shot at the memory bottleneck + energy efficient chips. Gigascale is back in for the $720M Series C. Congrats @goodwin_ml! They’re hiring: fractile.com/jobs

We’re delighted to announce that Fractile has raised $720M in a Series C led by @thrivecapital @foundersfund @lightspeedvp and @redpoint, with further participation by brilliant new and existing investors.

Post attachment
Mike Schroepfer avatar
Mike Schroepfer@schrep·Nov 6

Building power at internet scale and speed!

We've been marking major milestones at @heronpower the past few weeks, from completing our first two prototype Heron Links to pushing power at 34.5kVAC in our backyard test facility in Scotts Valley. Come join the fun, we're hiring!

Post attachmentPost attachmentPost attachment
Mike Schroepfer avatar
Mike Schroepfer@schrep·Nov 5

Hyundai Ioniq 5 Waymo spotted testing in Palo Alto!

0:05
Mike Schroepfer avatar
Mike Schroepfer@schrep·Nov 5

No renders just turbo machinery

Lots of $$ going into turbine startups and more renders to go with them. Meanwhile Arbor (@ArborEnergyCo) is testing their 1MW sCO2 system and scaling up to 25MW in 2027. Apply if you want to build time-to-power with a bunch of former SpaceX and GE engineers.

Post attachmentPost attachmentPost attachment
Mike Schroepfer avatar
Mike Schroepfer@schrep·Nov 5

Keep finding obvious alpha in applying 27 years of lessons building internet infra to building out our power grid. Continuous monitoring, buffering, software defined networking. All coming to your power grid.

Alicia Harmon avatar
⟳ Mike Schroepfer retweeted
Alicia Harmon@Alicia_AERC·Nov 3

Sunday thoughts reading this piece, nursing a candy hangover: “When I interview machinists and engineers who work on the factory floor, the word “muscle memory” comes up a lot. It might be the angle used to hold a tool, the timing of an operation or an adjustment for a particular machine. This is the tacit knowledge that physical AI needs to collect. And it won't be collected through data farms like robotics data.” In the recent Field Trips episode with @schrep there’s a back-and-forth about how important the senior experience on our team is, and how there is just something in the gut simulations can’t capture. And the relationships we have with suppliers / ‘guys who know a guy’ is another point that comes up a few times. It’s an exciting time to be thinking about building new machines and the manufacturing capability behind them. But it does strike me that the special use cases are ones with a human in the loop, technology or systems that make expertise exponential. The automation/replacement plays feel a longgggg way off. But I digress… What is cooler IMHO is seeing the way the ecosystem is rippling outward in the wake of the hardware renaissance in SoCal, making the human reality of what makes Shenzhen’s manufacturing capability so robust. We can have the advanced technology and the vertical integration and learning acceleration AND an edge that stems from deep human relationships and a respect for every partner in the process. Link to Field Trips Video youtu.be/yB9-hbGGBdk?is=VoEM…

Michelle Sun @michellelsun Oct 27 Article Can America Leapfrog Shenzhen? To rebuild a domestic manufacturing base in the US, we first need to understand what we’re trying to replicate. My visits to factories in Shenzhen and conversations with US builders have made me

Rhoda AI avatar
⟳ Mike Schroepfer retweeted
Rhoda AI@RhodaAI·Oct 29

Robots rarely fail on the clean runs. They fail on the edge cases. On Field Trips, @schrep and @gigascale watched our robot hit one live on camera and work through it on its own, with no human stepping in. That's what out DVA model is built for. 🎥 youtu.be/VnzfHhovGDk

0:30
Radiant avatar
⟳ Mike Schroepfer retweeted
Radiant@RadiantNuclear·Oct 29

Radiant has secured U.S. Department of Energy (@ENERGY) approval of its Documented Safety Analysis for Kaleidos, making it the only microreactor to be reviewed and approved by the DOE for full power operations. The approval is not just for criticality, but for Radiant’s full power, full temperature testing at the Idaho National Laboratory. Radiant will test the complete reactor system using the same design it will manufacture at scale. Next up in the path to commercialization: Readiness Review. Read more here 🔗 radiantnuclear.com/news/radi…

0:05
Mike Schroepfer avatar
Mike Schroepfer@schrep·Oct 28

Yo @SlackHQ don't pollute my notifs with a Trick Or Treat Halloween theme upsell thingy. You are a work tool and I triage my bolds with extreme prejudice. kthnxbye