Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Today, we're introducing AC2, the Applied Compute Agent Cloud, to enable every team to train, serve, and improve their own frontier models.

Applied Compute

8,255 subscribers

206,702 views • 1 month ago •via X (Twitter)

Science & Technology Education

Anya Rossi• Live Now

Private livecam show

33 Comments

Applied Compute1 month ago

AC2 turns post-training into a repeatable, engineered process. Use your own harness to train up to multi-trillion-parameter models with all the latest research techniques.

Applied Compute1 month ago

Ari, our research agent, monitors training, reads traces to find failure modes, analyzes results, and turns insights into better data, graders, and experiments.

Applied Compute1 month ago

When you have the right checkpoint, deploy it to production with AC2 inference: dedicated, low latency, and 99.9% uptime all while preserving the configuration used during training.

Applied Compute1 month ago

The most useful data comes from real user interactions, feedback, and traces. With self-distillation, your model can learn from production traces even when the original environment can’t be replayed.

Applied Compute1 month ago

Over the last six months, AC2 has powered our Applied AI team’s work for Microsoft, NVIDIA, Cognition, Harvey, DoorDash, and others. Now, we’re excited to empower every AI team with their own model factory. AC2 is available in private beta. Reach out to learn more.

Ishan Mukherjee1 month ago

Congrats @rhythmrg @ypatil125 @lindensli @michaelzchen5 and team! Looking forward to working as we post train our @rox_ai models.

John Schulman1 month ago

congrats!

vincent sunn chen1 month ago

congrats @ypatil125 @lindensli @michaelzchen5 and team!!

Apoorv Agrawal1 month ago

Let's go!! @ypatil125 @lindensli Agastya and team!

Dan Biderman1 month ago

congrats guys!

Jonathan Jacobi1 month ago

👑

Francesca LaBianca1 month ago

Congrats team!

andrew1 month ago

congrats on the launch!

Jared Rosner1 month ago

🔥

kevin1 month ago

congrats on the launch AC team!

Maanav Khaitan1 month ago

congrats on the launch!

Lina T1 month ago

the team is cooooking 🧑‍🍳

Zongheng Yang1 month ago

Congrats @ypatil125 @lindensli and team, cooking 🔥

Ramin1 month ago

great work! can I get access?

Neeraj Kumar1 month ago

Improvement loops need reversible deployment.

Shriram Sridharan1 month ago

Congrats on the launch !!

Praneet1 month ago

Incredible launch, congrats @rhythmrg and team!

angela zhang1 month ago

So exciting!! Let’s go!!

Omeed Tavakoli1 month ago

Wow

Ishan Gupta1 month ago

Congrats on the launch!

Alex1 month ago

Awesome !! This gives companies more control over their models. Something awesome is happening ! We just need to combine it all...!!

angel hernandez1 month ago

What happens when the us government regulates frontier models and we are only allowed to use claude

Sebastian Tan1 month ago

insane

Sebastian Buzdugan1 month ago

most teams lack enough clean data for frontier training, how does ac2 prevent garbage feedback loops

jian1 month ago

It’s like Tinker

Rompel1 month ago

"Enable every team to train frontier models" is the pitch every MLOps startup makes right before the pricing page shows you still need a stack of H100s and an RL infra team to use it. What's actually managed here — the training loop, or just the serving layer around someone else's weights?

Robby Manihani1 month ago

congrats @ypatil125 and team!

刘朝 Zhao Liu1 month ago

The trace-to-checkpoint loop is the real product: research agents find failure modes, graders turn them into evidence, and deployment preserves the tested configuration. Account and session state belong in that trace too:

Related Videos

Today, we’re excited to launch the Datology Curation Studio, customizable frontier data curation for every team training its own models. Data quality is the ultimate compute multiplier, and Curation Studio makes it easy for everyone.

Today, we’re excited to launch the Datology Curation Studio, customizable frontier data curation for every team training its own models. Data quality is the ultimate compute multiplier, and Curation Studio makes it easy for everyone.

DatologyAI

4,270,178 views • 3 days ago

With the new MAI models and Frontier Tuning capabilities we announced today, we're focused on helping every company move from just consuming a frontier model to fully participating at the frontier.

With the new MAI models and Frontier Tuning capabilities we announced today, we're focused on helping every company move from just consuming a frontier model to fully participating at the frontier.

Satya Nadella

101,187 views • 4 months ago

Enterprise AI deployments today are frozen in time. Model capabilities stagnate in production. The problem compounds because companies aren’t static either. Every time your company improves, the model falls further behind. The bottleneck is continual learning. How does a model do something once and improve from feedback? The future of enterprise AI is Specific Intelligence: custom models teams own, trained on a company’s choices, interaction by interaction, using internal knowledge general models cannot access. Applied Compute helps companies train, serve, own, and improve custom models. Thanks Apoorv Agrawal for having Yash Patil at MS&E 435 to talk about the future of model training.

Enterprise AI deployments today are frozen in time. Model capabilities stagnate in production. The problem compounds because companies aren’t static either. Every time your company improves, the model falls further behind. The bottleneck is continual learning. How does a model do something once and improve from feedback? The future of enterprise AI is Specific Intelligence: custom models teams own, trained on a company’s choices, interaction by interaction, using internal knowledge general models cannot access. Applied Compute helps companies train, serve, own, and improve custom models. Thanks Apoorv Agrawal for having Yash Patil at MS&E 435 to talk about the future of model training.

Applied Compute

12,871 views • 4 months ago

Today we're introducing Cloud Sandboxes: same microVM isolation as Docker Sandboxes on your laptop, same CLI, always-on Docker-managed compute. Close your laptop. Keep the agent running. Move between local and cloud with one command. $250 credit for new accounts:

Today we're introducing Cloud Sandboxes: same microVM isolation as Docker Sandboxes on your laptop, same CLI, always-on Docker-managed compute. Close your laptop. Keep the agent running. Move between local and cloud with one command. $250 credit for new accounts:

Docker

105,939 views • 18 days ago

Introducing Lab: A full-stack platform for training your own agentic models Build, evaluate and train on your own environments at scale without managing the underlying infrastructure. Giving everyone their own frontier AI lab.

Introducing Lab: A full-stack platform for training your own agentic models Build, evaluate and train on your own environments at scale without managing the underlying infrastructure. Giving everyone their own frontier AI lab.

Prime Intellect

3,155,803 views • 8 months ago

Following this week's frontier video model launches, we're opening up early access to something we've been building. Introducing fal Agent. A creative partner for working across the best image, video, and 3D models.

Following this week's frontier video model launches, we're opening up early access to something we've been building. Introducing fal Agent. A creative partner for working across the best image, video, and 3D models.

fal

104,325 views • 2 months ago

Runway 🤝 Vera Rubin Excited to continue working with NVIDIA on frontier models and frontier compute

Runway 🤝 Vera Rubin Excited to continue working with NVIDIA on frontier models and frontier compute

Cristóbal Valenzuela

12,309 views • 7 months ago

Today we're incredibly excited to announce Dreamscale Labs (YC F26) (YC F26)! We're an applied research company putting robot brains in the cloud. Robot AI models are growing larger and on-board compute is getting increasingly inaccessible due to cost. We're taking a contrarian bet that the future is moving inference off the robot and into the cloud, while still meeting real-time deadlines required by robot motion. We're building the real-time inference infrastructure that enables frontier physical AI models to run smoothly on your robot. We optimize for low latency across the full stack: from the model, down to the GPU kernels, and over the network. Our mission is to unlock a future of safe and intelligent robots doing meaningful work, and we think this is only possible by letting you deploy and run robot models in the cloud. Reach out to us at founders@dreamscalelabs.com or visit our website at to learn more. Kevin L Antoine Nguyen Eric R

Today we're incredibly excited to announce Dreamscale Labs (YC F26) (YC F26)! We're an applied research company putting robot brains in the cloud. Robot AI models are growing larger and on-board compute is getting increasingly inaccessible due to cost. We're taking a contrarian bet that the future is moving inference off the robot and into the cloud, while still meeting real-time deadlines required by robot motion. We're building the real-time inference infrastructure that enables frontier physical AI models to run smoothly on your robot. We optimize for low latency across the full stack: from the model, down to the GPU kernels, and over the network. Our mission is to unlock a future of safe and intelligent robots doing meaningful work, and we think this is only possible by letting you deploy and run robot models in the cloud. Reach out to us at [email protected] or visit our website at to learn more. Kevin L Antoine Nguyen Eric R

Chris Yoo

48,653 views • 25 days ago

Introducing Odyssey-2 Pro—a frontier world model that generates long-running, interactive simulations in 720p! We're also launching the first world model API, to enable devs to build magical apps. We're now in the GPT-2 era of world models. Let the explosion of apps commence!

Introducing Odyssey-2 Pro—a frontier world model that generates long-running, interactive simulations in 720p! We're also launching the first world model API, to enable devs to build magical apps. We're now in the GPT-2 era of world models. Let the explosion of apps commence!

Odyssey

166,651 views • 8 months ago

Today, we’re introducing Decide Agent, a frontier-level AI agent for Excel analysis. Decide Agent scored 82.50% on SpreadsheetBench, the public standard benchmark for Excel AI agents, ranking #4 globally. This is the same leaderboard used by Claude, OpenAI, and Microsoft to evaluate their spreadsheet agents. We’re here to push the frontier in Excel and data analytics. Decide Agent is now generally available at trydecide . ai

Today, we’re introducing Decide Agent, a frontier-level AI agent for Excel analysis. Decide Agent scored 82.50% on SpreadsheetBench, the public standard benchmark for Excel AI agents, ranking #4 globally. This is the same leaderboard used by Claude, OpenAI, and Microsoft to evaluate their spreadsheet agents. We’re here to push the frontier in Excel and data analytics. Decide Agent is now generally available at trydecide . ai

Ab.

40,512 views • 8 months ago

Greg Brockman on AI-designed chips and the future of compute >openai used their own models to design their chips, faster and probably better than human engineers >in the future everyone would have an agent that works 24/7 for them. Restriction: compute, at the moment

Greg Brockman on AI-designed chips and the future of compute >openai used their own models to design their chips, faster and probably better than human engineers >in the future everyone would have an agent that works 24/7 for them. Restriction: compute, at the moment

Chubby♨️

209,854 views • 1 year ago

Introducing Kled Labs. Every major AI model learned from the internet. We're betting on people instead. One goal: train our own models to beat closed models at important human tasks, using our data alone. All data is sourced from our data application layer and our network of 600,000+ contributors. We're hiring:

Introducing Kled Labs. Every major AI model learned from the internet. We're betting on people instead. One goal: train our own models to beat closed models at important human tasks, using our data alone. All data is sourced from our data application layer and our network of 600,000+ contributors. We're hiring:

Kled AI

31,728 views • 12 hours ago

Every AI response needs a GPU to run it. Every robot decision needs compute to power it. Every autonomous agent needs infrastructure to live on. We're building that infrastructure. ROBA Compute. Coming soon.

Every AI response needs a GPU to run it. Every robot decision needs compute to power it. Every autonomous agent needs infrastructure to live on. We're building that infrastructure. ROBA Compute. Coming soon.

Roba Labs

49,108 views • 4 months ago

We're launching Niteshift – the full-stack cloud for coding agents Verification is the new bottleneck. Software teams can now define their dev environment and verification tools once. Then run any frontier agent in the cloud: Claude Code, Codex, or OpenCode

We're launching Niteshift – the full-stack cloud for coding agents Verification is the new bottleneck. Software teams can now define their dev environment and verification tools once. Then run any frontier agent in the cloud: Claude Code, Codex, or OpenCode

Sajid Mehmood

102,791 views • 4 months ago

Introducing multi-agent orchestration in Oz, with support for Claude Code, Codex and the Warp Agent. Use /orchestrate to delegate complex tasks across a team of agents, locally or in the cloud.

Introducing multi-agent orchestration in Oz, with support for Claude Code, Codex and the Warp Agent. Use /orchestrate to delegate complex tasks across a team of agents, locally or in the cloud.

Warp

1,198,812 views • 4 months ago

Mercor connects experts in their field with leading AI labs to test, challenge, and improve frontier AI models. Whether you’re a software engineer, PhD, consultant, or doctor, you can earn up to $200/hr while working on your own terms. Apply today at

Mercor connects experts in their field with leading AI labs to test, challenge, and improve frontier AI models. Whether you’re a software engineer, PhD, consultant, or doctor, you can earn up to $200/hr while working on your own terms. Apply today at

Mercor

215,529 views • 1 year ago

Introducing runneth, the AI brain for Marketing. It's like having your own mac mini in the cloud. Seriously. We built infrastructure to give *every* customer their own dedicated virtual machine.

Introducing runneth, the AI brain for Marketing. It's like having your own mac mini in the cloud. Seriously. We built infrastructure to give every customer their own dedicated virtual machine.

Reza Khadjavi

14,976 views • 3 months ago

We're introducing hybrid compute for all users of the Perplexity Mac app. This will allow Computer to orchestrate local models that can run locally on Mac, particularly for agent steps involving sensitive and private files (eg your bloodwork, tax returns, litigation, etc).

We're introducing hybrid compute for all users of the Perplexity Mac app. This will allow Computer to orchestrate local models that can run locally on Mac, particularly for agent steps involving sensitive and private files (eg your bloodwork, tax returns, litigation, etc).

Aravind Srinivas

526,602 views • 1 month ago

“We want to orchestrate the world’s compute, in the same way Bitcoin did, but to train models that can rival ChatGPT” crux on the This Week in Startups podcast, sharing the motivations behind IOTA ・ SN9: to train frontier models efficiently and at scale, in an age where CapEx for centralized training is reaching unsustainable highs.

“We want to orchestrate the world’s compute, in the same way Bitcoin did, but to train models that can rival ChatGPT” crux on the This Week in Startups podcast, sharing the motivations behind IOTA ・ SN9: to train frontier models efficiently and at scale, in an age where CapEx for centralized training is reaching unsustainable highs.

Macrocosmos

14,545 views • 5 months ago

Your data is your edge, but only if your AI is built on it. Rent a generic model and so can your competitor. The companies with an edge are deploying custom models that they own and improve over time. Our co-founder Rhythm Garg recently stopped by to share how companies are owning their intelligence with Applied Compute.

Your data is your edge, but only if your AI is built on it. Rent a generic model and so can your competitor. The companies with an edge are deploying custom models that they own and improve over time. Our co-founder Rhythm Garg recently stopped by to share how companies are owning their intelligence with Applied Compute.

Applied Compute

119,306 views • 4 months ago