The next phase of AI adoption will depend on making intelligence more efficient, not just more capable. In this episode, Jaya is joined by Stanford Ph.D. researchers Jon Saad-Falcon and Avanika Narayan to discuss how local and hybrid AI can help bring these tools to billions more people.

Much of today’s LLM inference runs through frontier models in centralized data centers. But Jon and Avanika argue that many tasks can already be handled by smaller, open-weight models, including models running on local hardware. Their concept of “intelligence per watt” measures how much task performance a given model-and-hardware configuration delivers relative to the power it consumes.

They also discuss OpenJarvis and Minions, two projects that explore how AI systems can use local models when possible and turn to cloud models when more capability is needed.

From there, the conversation turns to what this shift could mean for OpenAI, Anthropic, and other frontier labs. As open-weight models improve and inference becomes cheaper and more distributed, Jon and Avanika argue that low switching costs and limited lock-in could put pressure on the labs’ margins. They also explain why enterprises should be careful about handing proprietary data and decision-making processes over to model providers.

They close with advice for Ph.D. students and academic researchers: rather than trying to reproduce the frontier labs’ work with fewer resources, choose problems where original ideas matter more than scale.

Intelligence per Watt: Measuring Intelligence Efficiency of Local AI - https://arxiv.org/abs/2511.07885

OpenJarvis: Personal AI, On Personal Devices - https://arxiv.org/abs/2605.17172

Minions: Cost-efficient Collaboration Between On-device and Cloud Language Models - https://arxiv.org/abs/2502.15964

In this episode, Joanne is joined by Azeem Azhar, founder of Exponential View, to unpack what it will take for AI to move from individual productivity gains to true org-level transformation.

The conversation builds on Joanne and Leo’s recent essay on the great reorg, which argues that AI’s full impact will only arrive when companies redesign how work gets done from first principles. 

Today, AI is already helping individuals work faster, but many organizations are struggling to translate that into gains at the team or company level. Azeem calls this problem “congestion.” When one part of the company speeds up, the next downstream process becomes the bottleneck. As AI takes on more work inside organizations, it reveals where they are too slow, too rigid, or too dependent on processes built around human constraints.

Joanne and Azeem explore what it will take to build AI-native companies, from new human roles like system architects, validators, and accountability owners to new tools designed for agents rather than humans.

For founders, Azeem argues that the most important metric may be cycle time: how quickly a company can learn from customers, ship, adapt, and repeat. The org chart is being redrawn, and the companies that thrive will be the ones that learn how to collaborate with agents instead of simply bolting AI to old workflows.

What we covered:

Links: 

Exponential View - https://www.exponentialview.co

The great org: A human’s guide - https://foundationcapital.com/ideas/the-great-reorg

Tennr co-founder and CEO Trey Holterman joins Joanne to talk about what it takes to make AI work in healthcare, and why the answer is not simply "better models."

Tennr focuses on patient flow: the process of getting patients from one provider to the next with the right records, insurance information, clinical context, and scheduling in place. It's a problem that sits in the messy middle of healthcare, where delays, denials, missing documentation, and manual workflows are the norm. Today, its platform processes referrals and intake documentation for roughly 10% of Americans each year.

Healthcare has long been considered too hard for startups to crack. But advances in LLMs have made it possible to build products that can handle the unstructured, edge-case-heavy nature of healthcare administration. As Trey explains, those advances have created an opening for founders willing to stay close to customers and learn the market directly.

Trey also gets into his first-time founder lessons and Tennr's path to scale. In the early days, he and his co-founders spent too much time building in isolation, then swung too far the other way by saying yes to nearly everything customers asked for. At one point, they were approaching $1M in revenue, only to realize they were spreading themselves across too many use cases. The decision to refocus resulted in a short-term hit to their revenue, but became the moment the team locked in and found product-market fit.

Along the way, Trey reflects on the discipline of questioning your own assumptions, how Tennr uses AI internally, and what the patient experience could look like if Tennr realizes its mission of making healthcare move faster.

Chapters:

Welcome to AI in the Real World! In this episode, Foundation Capital Partner Joanne Chen sits down with Ken Goldberg, professor of engineering at UC Berkeley and co-founder of Ambi Robotics, a company applying AI-enabled robotics to transform the logistics industry.

Ken has spent more than four decades working at the intersection of robotics and AI, focusing on one of the most persistent challenges in the field: how machines perceive and manipulate the physical world.

Ken shares why tasks that seem trivial to humans, like picking up a glass or folding laundry, remain profoundly difficult for robots, and why the physical world introduces a level of uncertainty that can't be fully simulated.

Our conversation also covers:

Chapters:

Welcome to AI in the Real World! In this episode, Foundation Capital Partner Jaya Gupta sits down with Karan Goel, co-founder and CEO of Cartesia, a company pioneering the use of state space models (SSMs) for AI applications.

Karan shared his journey from pursuing a PhD at Stanford to co-founding Cartesia with Albert Gu - taking their academic research on SSMs and turning it into a product that is initially focused on voice applications.

Karan explained how SSMs differ fundamentally from transformers, offering a more efficient, memory-based architecture that processes information sequentially rather than in batches. This approach enables more human-like AI systems that can remember and adapt to new information in real-time.

The conversation also covers:

Chapters:

Welcome to AI in the Real World! In this episode, Foundation Capital Partner Jaya Gupta sits down with Rohan Taori, a researcher on Anthropic's multimodal pre-training team.

Within the AI community, Rohan is best known for co-creating Alpaca, a project that demonstrated how fine-tuning Meta's LLaMA model could achieve ChatGPT-level performance for under $600.

Rohan shares his journey from early work in computer vision at UC Berkeley to his Ph.D. at Stanford, where he explored methods for making AI more accessible.

He explains the technical breakthroughs behind Alpaca, including self-instruct, a method that uses a stronger language model (OpenAI's text-davinci-003) to generate synthetic data that is then used to fine-tune a weaker model (Llama). This approach, which underpins Alpaca and its follow-up projects AlpacaFarm and AlpacaEval, illustrates how small-scale post-training can significantly enhance model performance.

The conversation also covers:

Chapters:

Welcome to AI in the Real World! In this episode, Foundation Capital Partner Sid Trivedi speaks to Chris "Tito" Sestito, the co-founder and CEO of HiddenLayer.

HiddenLayer is a cybersecurity platform purpose-built for AI systems. Tito's journey into AI security began at Cylance, where he led the response to one of the first real-world AI attacks. This experience led him to start HiddenLayer, which recently raised a $50M Series A.

In this conversation, Tito shares his insights into the evolving AI threat landscape, from shadow AI to the growing arsenal of strategies that bad actors are using to target generative models. He explains why data poisoning is one of the most urgent threats facing humanity, and why safeguarding AI will define the future of cybersecurity.

Learn more about HiddenLayer here: https://hiddenlayer.com

Chapters:

Welcome to AI in the Real World! In this episode, Foundation Capital General Partner Joanne Chen speaks with Melody Meckfessel, the CTO of Jasper.

Melody has been a pioneering technical leader and co-founder of two startups. She spent 15 years running engineering teams at Google, where she helped shape Google Cloud and improve search. Here, she shares her journey from the early days of search to today's cutting-edge work in generative AI for marketers at Jasper.

Welcome to AI in the Real World! I'm Andrew Han, a partner at Foundation Capital, and I'll be your host today. I work closely with technical founders who are building software to solve hard, valuable problems.

I'm excited to chat with Darius Lam, founder and CEO of NEX, a recent addition to our portfolio. Darius was previously head of product for computer vision at Cerebras, another Foundation Capital company that's building wafer-scale chips for AI workloads.

Today, Darius develops state-of-the-art multimodal foundation models at NEX. Our conversation explores the ins and outs of creating these models, including training efficiency, output controllability, and the last mile of AI deployment. I also get Darius's take on the evolution of open source AI and what every AI founder should know before they write their first line of code.

Chapters:

Welcome to AI in the Real World. I'm Steve Vassallo, a General Partner at Foundation Capital. In this series, we talk with leading AI researchers about their groundbreaking work and how it's being applied in real businesses today.

Joining me is Sean Lie, co-founder and CTO of Cerebras Systems. I've known Sean since early 2016 before leading the Series A investment in Cerebras. At the time, Sean was just wrapping up a three-year stint as Chief Architect of data center solutions at AMD, which he joined via AMD's acquisition of his last startup, SeaMicro.

Sean is one of the biggest brains I know around hardware systems, bringing a rare expertise and passion for advanced architectures, many of which break from tradition and traditional approaches to computing.

In this episode, Sean guides us through the world of AI hardware innovation. He breaks down the concept of wafer-scale processing, the challenges of building monstrous AI chips, and the massive potential of sparsity. He offers valuable lessons from his journey as a deep tech entrepreneur, including the importance of thinking end-to-end and planning for scale from day one.

We close by peering into AI's future and sharing advice for founders taking on similarly complex, system-level problems.