Open||Source||Data

From energy bottlenecks to proprietary GPU ecosystems, the CEO of TensorWave, Darrick Horton explains why today’s AI scale is unsustainable—and how open-source hardware, smarter networking, and nuclear power could be the fix.QUOTESDarrick Horton“The energy crisis is getting worse every day. It’s very hard to find data center capacity—especially capacity that can scale. Five years ago, 10 or 20 megawatts was considered state-of-the-art. Now, 20 is nothing. The real hyperscale AI players are looking at 100 megawatts minimum, going into the gigawatt territory. That’s more than many cities combined just to power one cluster.”Charna Parkey“We’re still training models in a very brute-force way—throwing the biggest datasets possible at the problem and hoping something useful emerges. That’s not sustainable. At some point, we have to shift toward smarter, more intentional training methods. We can’t afford to be wasteful at this scale.”TIMESTAMPS[00:00:00] Introduction[00:01:00] Founding TensorWave[00:04:00] AMD as a Viable Alternative[00:08:00] Open Source as a Startup Enabler[00:09:30] Launching ScalarLM[00:12:00] ScalarLM Impact and Reception[00:14:30] Roadmap for 2025[00:16:00] Technical Advantages of AMD[00:18:00] Emerging Open Source Infrastructure[00:20:00] Broader Societal Issues AI Must Address[00:22:00] AI’s Impact on Global Energy[00:26:00] Fundamental Hardware vs. Human Efficiency[00:30:00] Data Center Density Evolution[00:34:00] Advice to Founders and Tech Trends[00:38:00] AI Energy Challenges[00:44:00] AI’s Rapid Impact vs. Internet[00:46:00] Monopoly vs. Democratization in AI[00:50:00] Close to Season Wrap Discussion and Predictions

Join Charna Parkey as she welcomes Anastasia Stasenko, CEO and co-founder of pleias, through her unique journey from philosophy to building open-source, energy-efficient LLMs. Discover how pleias is revolutionizing the AI landscape by training models exclusively on open data and establishing a precedent for ethical and socially acceptable AI. Learn about the challenges and opportunities in creating multilingual models and contributing back to the open-source community. QUOTES[00:00:00] Introducing Anastasia and pleias[00:02:00] From Philosophy to AI[00:06:00] The Problem of Generic Models[00:10:00] Open Weights vs. Open Source vs. Open Science[00:14:00] Why Open Data Matters[00:18:00] High-Quality, Specialized Models[00:22:00] Multilingual Challenges[00:26:00] Global Inclusion Requires Small Models[00:30:00] Using and Contributing to Wikidata[00:38:00] The Future: Specialized Models[00:48:00] Advice for Newcomers[00:54:00] Cultural Sensitivity and Data Representation[00:50:00] Leo’s Takeaways[00:52:00] Charna on Ethical, Verifiable AI[00:54:00] Representation vs. Exclusion[00:56:00] Letting People Be More Human[00:57:30] Applied, Transformative AIQUOTESCharna:"If you didn’t make it represented in the data, then we’re leaving another culture behind... So which one are you wanting to do, misrepresent them or just completely leave them behind from this technical revolution?"Anastasia:"The real issue now is that the lack of diversity in the current AI labs leads to the situation where all LLMs look alike."Anastasia:"Being able to design, to find, and also to create the appropriate data mix for large language models is something that we shouldn't really forget about when we talk about the success of what large language models are."

Discover how Rackspace Spot is democratizing cloud infrastructure with an open-market, transparent option for cloud servers. Kevin Carter, Product Director at Rackspace Technology, discusses Rackspace Spot's hypothesis and the impact of an open marketplace for cloud resources. Discover how this novel approach is transforming the industry. TIMESTAMPS[00:00:00] – Introduction & Kevin Carter’s Background[00:02:00] – Journey to Rackspace and Open Source[00:04:00] – Engineering Culture and Pushing Boundaries[00:06:00] – Rackspace Spot and Market-Based Compute[00:08:00] – Cognitive vs. Technical Barriers in Cloud Adoption[00:10:00] – Tying Spot to OpenStack and Resource Scheduling[00:12:00] – Product Roadmap and Expansion of Spot[00:16:00] – Hardware Constraints and Power Consumption[00:18:00] – Scrappy Startups and Emerging Hardware Solutions[00:20:00] – Programming Languages for Accelerators (e.g., Mojo)[00:22:00] – Evolving Role of Software Engineers[00:24:00] – Importance of Collaboration and Communication[00:28:00] – Building Personal Networks Through Open Source[00:30:00] – The Power of Asking and Offering Help[00:34:00] – A Question No One Asks: Mentors[00:38:00] – The Power of Educators and Mentorship[00:40:00] – Rackspace’s OpenStack and Spot Ecosystem Strategy[00:42:00] – Open Source Communities to Join[00:44:00] – Simplifying Complex Systems[00:46:00] – Getting Started with Rackspace Spot and GitHub[00:48:00] – Human Skills in the Age of GenAI - Post Interview Conversation[00:54:00] – Processing Feedback with Emotional Intelligence[00:56:00] – Encouraging Inclusive and Clear Collaboration QUOTESCHARNA PARKEY“If you can’t engage with this infrastructure in a way that’s going to help you, then I guarantee you it’s not up to par for the direction that we’re going. [...] This democratization — if you don’t know how to use it — it’s not doing its job.”KEVIN CARTER“Those scrappy startups are going to be the ones that solve it. They’re going to figure out new and interesting ways to leverage instructions. [...] You’re going to see a push from them into the hardware manufacturers to enhance workloads on FPGAs, leveraging AVX 512 instruction sets that are historically on CPU silicon, not on a GPU.”

Pete Pachal, founder of The Media Copilot, dives into the transformative role of AI in media and journalism. With over two decades in technology reporting, he discusses how journalists can harness AI as a tool for innovation. Pachal emphasizes the importance of maintaining creative control over AI outputs and explores the evolving dynamics between AI-generated content and audience trust. He also reflects on the challenges faced by traditional journalism in adapting to new technologies while preserving integrity in storytelling.

In this episode, Dr. Joan Bajorek—AI entrepreneur, author of Your AI Roadmap, and founder of Clarity AI—joins Charna Parkey to talk about what it really takes to build a future in AI. From career pivots and layoff anxiety to financial transparency and finding joy in your work, Joan shares practical advice and personal stories navigating fear, burnout, and career uncertainty in tech, while staying grounded in purpose, community, and long-term resilience.TIMESTAMPS[00:00:00] — Introduction to Joan Bajorek & Her Work[00:02:00] — Transparency About Finances and Career[00:04:00] — The Taboo Around Talking About Money[00:06:00] — Resilience During Tech Layoffs[00:08:00] — How to Get Credit for Your Work[00:12:00] — Should You Chase an AI Job?[00:14:00] — Career Goals vs. Financial Security[00:16:00] — Translating Academic and Life Skills into Tech[00:18:00] — Defining and Finding Joy in Work[00:20:00] — Multiple Income Streams and Personal Freedom[00:24:00] — AI’s Near-Future Impact on Jobs and Industries[00:26:00] — Data and AI Opportunities in Underexplored Domains[00:34:00] — Creating Scalable, Alternative Income Models[00:36:00] — How Joan Maintains Long-Term Motivation[00:42:00] — Post-Interview DiscussionQUOTESJoan Bajorek"Networking is how I've gotten the best opportunities and jobs of my life... LinkedIn has this research about how after COVID layoffs, 70% of people landed their next job based on an intro."Charna Parkey"I always try to strive for transparency, and I get such mixed results where at work with coworkers, it's absolutely valued. And then there seems to always be some sort of consequences in my personal life."

Dr. Jason Corso joins Charna Parkey to debate the critical role of data quality, how its transparency shapes AI development and the rise of smaller, domain-specific AI models - making 2025 the year of small, specialized AI. QUOTESCharna Parkey"Knowing the right data is incredibly important, because it'll save you money, but predicting the impact of that data means that you don't have to do the training at all to even directionally know if it's going to work out, right?"Jason Corso "You can't understand and analyze an AI system in the way you can analyze open source software if you don't have access to the data."Timestamps[00:00:00] - Introduction[00:02:00] - Jason Corso’s journey on open source[00:08:00] - The importance of data in AI[00:10:00] - Voxel 51's mission[00:14:00] - The value of open source and the importance of data in AI systems[00:20:00] - Recent discoveries in AI[00:28:00] - The cost of training AI models[00:36:00] - Cooperative AI in healthcare[00:40:00] - Charna Parkey on the impact of AI in education[00:56:00] -The year of small AI 

In this episode of Open Source Data, Charna Parkey talks with Alex Gallego, CEO and founder of Redpanda Data, about his journey as a builder, the evolution of Redpanda, and the company's new agent framework for the enterprise. Alex shares insights on low-latency storage, distributed stream processing, and the importance of developer experience to the growth of AI and the Open Source space. Timestamps[00:00:00] Introduction[00:02:00] Alex Gallego talks about his background[00:04:00] Charna Parkey discusses the importance of hands-on experience in learning.[00:06:00] Alex explains the origins of Red Panda and how it emerged from challenges in the streaming space.[00:08:00] Alex details the evolution of Red Panda, its use of C-Star and FlatBuffers, and its low-latency design.[00:11:00] Alex discusses the positioning of Kafka versus Red Panda in the market.[00:20:00] Alex introduces Red Panda's new agent framework and multi-agent orchestration.[00:24:00] Alex explains how Red Panda fits into the evolving landscape of AI-powered applications.[00:30:00] The future of multi-agent orchestration.[00:44:00] Thoughts on AI model training and data retention.[00:46:00] Alex encourages future founders and shares his perspective on risk-taking.[00:50:00] Charna Parkey and Leo Godoy discuss the key takeaways from the conversation with Alex Gallego.[00:52:00] Charna reflects on open source trends and the role of developer experience in adoption.[00:54:00] Charna and Leo talk about the different types of founder journeys and the importance of team dynamQuotes Charna Parkey"For AI, unifying historical and real-time data is critical. If you're just using nightly or monthly data, it doesn’t match the context in which your prediction is being made. So it becomes very important in the future of applying AI because you need to align those things."Alex Gallego"Every app is going to span three layers. The first layer is going to be your operational layer, just like you have to do business right now. Then there always has to be an analytical layer, and the third layer is this layer of autonomy."

In this episode, we dive deep into the world of neuro-symbolic AI with Emin Can Turan, CEO of Pebbles AI. Learn how this technology combines neuroscience, behavioral economics, and AI to revolutionize B2B go-to-market strategies. Emin explains how neuro-symbolic AI bridges the gap between human logic and machine learning, enabling smarter, context-aware systems that democratize complex workflows for startups and enterprises alike.Timestamps[00:00:00] - Introduction by Charna Parkey and introduction of Emin Can Turan.[00:02:00] - Emin’s journey to AI and his background in go-to-market strategies.[00:06:00] - Emin explains his deep R&D phase and the development of neuro-symbolic AI.[00:08:00] - Emin describes the architecture of their AI system, including neuro-symbolic AI, generative AI, and agentic frameworks.[00:10:00] - Explanation of neuro-symbolic AI and its relevance to domain-specific problems.[00:12:00] - Discussion on the components of go-to-market strategies and the role of psychology and communication.[00:16:00] -The limitations of generative AI and how they applied strict communication tactics.[00:22:00] - Discussion on the importance of contextual science and data insights.[00:24:00] - The three agentic frameworks they use in their system.[00:26:00] - Explanation of how users control the product and the two co-pilots (strategy and execution).[00:36:00] - The ethical implications of AI and the potential for misuse.[00:38:00] - Discussion on the future of AI and the balance between dystopian and hopeful outcomes.[00:40:00] - Emin emphasizes the importance of truth and transparency in AI development.[00:42:00] - Emin shares his personal motivation for building his AI startup.[00:48:00] - Closing remarks and discussion on the user experience of their platform.[00:50:00] - Charna and Leo discuss the connection between Emin's work and the open-source community.QuotesEmin Can Turan"I felt that this was the future and that AI was the only technology that can digitalize this level of complexity for everyone to use. Nothing else could, you know, you can't use normal neural networks to do this. Even generative AI is not sufficient enough."Charna ParkeyI would love to be able to use Gen AI for more personal things. I love technology. I have the Oura Ring. I've got the Apple Watch. I want to feed that data into something that can somehow tell me and others, here's your state of mind. Here's what you're going to be affected by. 

Suzanne El-Moursi, Co-founder and CEO of BrightHive, shares her journey from corporate life to entrepreneurship. She discusses how AI, specifically her platform BrightBot, empowers non-technical teams to harness data insights effectively. The conversation delves into the challenges of data management for those without technical backgrounds and highlights the transformative role of open-source solutions. Suzanne also emphasizes the importance of community in fostering innovation and the future of AI in enhancing workplace experiences and education.

Kicking off Open Source Data Season 7, Charna Parkey welcomes the CEO and Founder of Invoke, Kent Keirsey to discuss his thoughts on licensing, copyright in generative AI, and the role of communities in building ethical, free-to-use technologies that can democratize technology and inspire global innovation.QuotesKent Keirsey "When we look at open source models, if you just release the weights, and you don't really release information on how the data set was captioned, for example, or how you construct the data set, if you don't really know how it got to the artifact that was released, as a user, you do not understand how it works."Charna Parkey But there's still a lot of claims by big tech right now about how anything on the internet should be fair use for training, even if, you know, it might have its own kind of copyrightTimestamps[00:02:00] - Kent Keirsey on his journey to open source[00:06:00] - Kent Keirsey on the Open Model Initiative (OMI)[00:08:00] -What makes a model truly open source[00:12:00] - The legal landscape of AI and copyright[00:14:00] - Kent Keirsey on the ethical implications of AI training data fair and use and AI development[00:26:00] Creativity, AI tools, personal AI models and recommendation algorithms:[00:32:00] - Kent Keirsey on TikTok and cultural clash:[00:38:00] - AI, self-reflection and a decision-making tool[00:42:00] - The Bria AI partnership[00:52:00] - The future of creativity, AI and Robotics:[01:00:00] - Final thoughts with producer Leo GodoyConnect with Kent KeirseyConnect with Charna Parkey