Modal's banner
Modal's profile picture

Modal

@modal • 39,340 subscribers

AI infrastructure that developers love 💚 Run inference, sandboxes, batch processing, training, and many other things on Modal

Shorts

Training large models with RL typically runs on a fixed homogeneous cluster of the latest accelerators for numerical stability and faster throughput. This throws away inference optimization work done for other vendors and generations, limits the capacity you can utilize, and is unnecessarily expensive. So we trained a model on agentic coding tasks using fully asynchronous RL and mixed-precision, mixed-accelerator samplers. With off-policy correction, training stayed stable through convergence, on par with homogeneous samplers, at a lower cost.

Training large models with RL typically runs on a fixed homogeneous cluster of the latest accelerators for numerical stability and faster throughput. This throws away inference optimization work done for other vendors and generations, limits the capacity you can utilize, and is unnecessarily expensive. So we trained a model on agentic coding tasks using fully asynchronous RL and mixed-precision, mixed-accelerator samplers. With off-policy correction, training stayed stable through convergence, on par with homogeneous samplers, at a lower cost.

160,884 views

From frogs to the frontier, Scott Wu covered a lot at Runtime: → Agents are moving from tab completion to “virtual employees” that own outcomes → Frontier models may only need ~2% of workloads; routing handles the rest → 91% of Cognition’s internal Devin sessions start without a human

From frogs to the frontier, Scott Wu covered a lot at Runtime: → Agents are moving from tab completion to “virtual employees” that own outcomes → Frontier models may only need ~2% of workloads; routing handles the rest → 91% of Cognition’s internal Devin sessions start without a human

46,788 views

We're expanding across Europe. This summer we moved into a bigger Stockholm office, and today we're adding London. We're hiring across GTM and engineering! Read more:

We're expanding across Europe. This summer we moved into a bigger Stockholm office, and today we're adding London. We're hiring across GTM and engineering! Read more:

92,415 views

Runtime speaker lineup is live! We're bringing together experts covering AI infrastructure, applications of AI in science and robotics, the future of software engineering, and more.

Runtime speaker lineup is live! We're bringing together experts covering AI infrastructure, applications of AI in science and robotics, the future of software engineering, and more.

63,177 views

Modal is proud to now support Claude Managed Agents with Modal Sandboxes.

Modal is proud to now support Claude Managed Agents with Modal Sandboxes.

95,726 views

Runtime is happening in two days. Join us to hear speakers from Anthropic, TypeSafe AI, NVIDIA, Ramp, and more.

Runtime is happening in two days. Join us to hear speakers from Anthropic, TypeSafe AI, NVIDIA, Ramp, and more.

13,876 views

Registration is live for Runtime. Apply to attend our inaugural conference for engineers running AI in production. October 1st live at The Midway in San Francisco.

Registration is live for Runtime. Apply to attend our inaugural conference for engineers running AI in production. October 1st live at The Midway in San Francisco.

39,417 views

The team at Runway is pushing the frontier of what's possible in world models. We're proud to work with them to power real-time inference for Runway Characters.

The team at Runway is pushing the frontier of what's possible in world models. We're proud to work with them to power real-time inference for Runway Characters.

68,896 views

We're bringing together our friends and community to celebrate our Series C. Join us at Noguchi's Sunken Garden in NYC on June 16th or at the Legion of Honor in SF on June 25th. Invites are limited, apply here:

We're bringing together our friends and community to celebrate our Series C. Join us at Noguchi's Sunken Garden in NYC on June 16th or at the Legion of Honor in SF on June 25th. Invites are limited, apply here:

44,139 views

Co-founder and CEO of adaption, Sara Hooker, is joining us on stage at Runtime. She leads a team of AI researchers and engineers building systems that are radically efficient and adaptable for intelligence, designed to keep evolving. She'll share how Adaption is rethinking what it means for AI to stay current. Apply to attend.

Co-founder and CEO of adaption, Sara Hooker, is joining us on stage at Runtime. She leads a team of AI researchers and engineers building systems that are radically efficient and adaptable for intelligence, designed to keep evolving. She'll share how Adaption is rethinking what it means for AI to stay current. Apply to attend.

11,754 views

To best support its merchants, DoorDash is building an intelligence system with a suite of AI agents. From Andy Fang:

To best support its merchants, DoorDash is building an intelligence system with a suite of AI agents. From Andy Fang:

15,860 views

We're excited to power LemonSlice's new video agent product. Turn your real-time voice agent into a video agent using a single photo. The underlying model—which autoscales on Modal 😉—streams infinite-length talking video at 20fps using a single GPU.

We're excited to power LemonSlice's new video agent product. Turn your real-time voice agent into a video agent using a single photo. The underlying model—which autoscales on Modal 😉—streams infinite-length talking video at 20fps using a single GPU.

17,225 views

Sora is INSANE. This is the video it generated for my prompt "The best serverless data/AI/ML infrastructure platform tool in the world" 🤯🤯🤯

Sora is INSANE. This is the video it generated for my prompt "The best serverless data/AI/ML infrastructure platform tool in the world" 🤯🤯🤯

32,512 views

We're excited to announce Tidbyt is joining Modal! 🎉

We're excited to announce Tidbyt is joining Modal! 🎉

23,931 views

Videos

LIVE
1.2k

Anya Rossi

sweetdream.ai

SweetDream.aiSponsored•Livecam
Streaming Now

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

HD live stream
Exclusive private shows
1.2k viewers online

Free preview available • Premium content

Today we're announcing that Modal has crossed $1B in annualized run rate, as well as raised a $750M Series D at a $15.75B valuation, led by Accel. Since our Series C, usage has tripled and we’ve grown more than 15x year-over-year. We are deeply grateful to the developers and companies building on Modal. There's still so much more to build, and we can't think of a better time to be building software.
1:41
modal's profile picture

Today we're announcing that Modal has crossed $1B in annualized run rate, as well as raised a $750M Series D at a $15.75B valuation, led by Accel. Since our Series C, usage has tripled and we’ve grown more than 15x year-over-year. We are deeply grateful to the developers and companies building on Modal. There's still so much more to build, and we can't think of a better time to be building software.

Modal

300,225 views • 4 days ago

The full Runtime agenda is live, including talks from: Scott Wu, Co-founder & CEO Cognition Diogo Almeida, Co-founder & CEO TypeSafe AI Dylan Patel, Founder & CEO SemiAnalysis Sara Hooker, Co-founder & CEO Adaption Alex Ratner, Co-founder & CEO Snorkel AI
0:26
modal's profile picture

The full Runtime agenda is live, including talks from: Scott Wu, Co-founder & CEO Cognition Diogo Almeida, Co-founder & CEO TypeSafe AI Dylan Patel, Founder & CEO SemiAnalysis Sara Hooker, Co-founder & CEO Adaption Alex Ratner, Co-founder & CEO Snorkel AI

Modal

68,742 views • 25 days ago

It is not too late to _actually_ own your inference. Introducing: Modal Auto Endpoints.
0:52
modal's profile picture

It is not too late to _actually_ own your inference. Introducing: Modal Auto Endpoints.

Modal

266,594 views • 3 months ago

What happens to code reviews, costs, and hiring when agents write most of the code? At Runtime, cat (Claude Code) and dax (OpenCode) got into: → Token spend at 1 to 2x payroll → Turning every incident into a code review eval → Why “way too many people” use the terminal → Hiring for product taste over 10x execution
0:23
modal's profile picture

What happens to code reviews, costs, and hiring when agents write most of the code? At Runtime, cat (Claude Code) and dax (OpenCode) got into: → Token spend at 1 to 2x payroll → Turning every incident into a code review eval → Why “way too many people” use the terminal → Hiring for product taste over 10x execution

Modal

11,173 views • 5 days ago

The Runtime keynote is live. Updates to Inference, Training, and Sandboxes including: 11:20 — Endpoint Candidates 16:30 — Modal Clusters 21:48 — VM Sandboxes 22:40 — Sandbox Sidecars And more
24:02
modal's profile picture

The Runtime keynote is live. Updates to Inference, Training, and Sandboxes including: 11:20 — Endpoint Candidates 16:30 — Modal Clusters 21:48 — VM Sandboxes 22:40 — Sandbox Sidecars And more

Modal

15,731 views • 13 days ago

The future of artificial intelligence is physical. Physical Intelligence runs robotic control inference on Modal with >2x lower latency than the lag between your brain and your finger.
0:20
modal's profile picture

The future of artificial intelligence is physical. Physical Intelligence runs robotic control inference on Modal with >2x lower latency than the lag between your brain and your finger.

Modal

95,966 views • 6 months ago

Unstick your AI Modal x KRAZAM
1:15
modal's profile picture

Unstick your AI Modal x KRAZAM

Modal

68,746 views • 6 months ago

GLM-5, the latest frontier open model from Z.ai, is available now on Modal. We partnered with to release an endpoint that will be free for a limited time.
0:45
modal's profile picture

GLM-5, the latest frontier open model from Z.ai, is available now on Modal. We partnered with to release an endpoint that will be free for a limited time.

Modal

58,785 views • 8 months ago

We sat down with Ramp to talk about their internal coding agent, Ramp Inspect, that now writes over half of their merged PRs. - Runs on Modal Sandboxes. - Full dev environments spin up in seconds. - Hundreds of sessions in parallel.
0:35
modal's profile picture

We sat down with Ramp to talk about their internal coding agent, Ramp Inspect, that now writes over half of their merged PRs. - Runs on Modal Sandboxes. - Full dev environments spin up in seconds. - Hundreds of sessions in parallel.

Modal

32,567 views • 8 months ago

Build a conversational voice bot with 1 second voice-to-voice latency with Modal, Pipecat AI, and open models. Modal works seamlessly with WebRTC, WebSockets, and tunneling to squash latency to an absolute minimum.
1:01
modal's profile picture

Build a conversational voice bot with 1 second voice-to-voice latency with Modal, Pipecat AI, and open models. Modal works seamlessly with WebRTC, WebSockets, and tunneling to squash latency to an absolute minimum.

Modal

29,559 views • 11 months ago

No more content to load