Ahuman-centeredAIfluencycompany.
We believe AI should amplify human dignity, not replace it. Our first commercial proof point is Academy: a practical way for thoughtful professionals to use AI with confidence, clarity, and judgment. Stories feeds the work; Studio and Ventures compound only as real needs emerge.
Listen. Teach.
Build. Back.
We are borrowing the shape of a modern AI-native media company, but sequencing it carefully. Academy leads now. Stories supplies the signal. Studio and Ventures come later as proof compounds.
Listen
Stories and audience conversations reveal where people are confused, what they are trying to do with AI, and where human judgment still matters most.
Teach
Academy is the launch focus. It turns repeated questions into a practical AI fluency framework: Orient, Ask, Synthesize, Check, and Apply.
Build
Studio stays exploratory until Academy and Stories reveal repeated workflow pain points that are worth turning into templates, tools, or lightweight products.
Back
Ventures remains future-facing until advisory, equity, or startup activity is real. The ambition is to support aligned builders only when the ecosystem has earned that role.
Our core values
These aren't aspirations—they're commitments that shape every conversation we have and every product we build.
AI should raise human dignity, not erode it
Technology should make people more capable, more creative, and more connected—never a replacement for the irreplaceable work of being human.
Curiosity over certainty
We ask questions more than we give answers. The best conversations leave room for uncertainty, complexity, and changing your mind.
People before products
Every product starts with a human problem worth solving. Technical capability is never the starting point—human flourishing is.
Build in public, iterate with humility
We share our experiments openly, gather feedback honestly, and change course when needed. Perfection is the enemy of progress.
Alex Chao
UC Berkeley • Harvard Business School
Alex Chao is the Founder & CEO of Humans of AI — an ecosystem built around the conviction that AI should amplify human dignity, not replace it. He's spent the last decade building AI systems at the frontier of what's possible, and asking whether what's possible is what's needed.
Currently at ByteDance, Alex leads the US Exploration team (12 people) as Managing Director of US Research, working on Seed's frontier AI models from generation 1.6 to Seed 2.0. His work spans novel evaluation frameworks for LLM accuracy and reliability, cross-geography human-in-the-loop data operations targeting deep research and multimodal benchmarks, and orchestrating global task forces that drove model performance on Humanity's Last Exam to near state-of-the-art.
Previously at Microsoft, Alex was Product Lead for Generative AI incubations in the Office of the CTO, where he led focused teams of 10+ engineers, researchers, and product managers shipping state-of-the-art AI techniques across Microsoft and the Fortune 100. He incubated and open-sourced three of Microsoft's most impactful AI projects—GraphRAG (32k GitHub stars), Semantic Kernel (27k stars), and AutoGen (54k stars)—pioneering agentic and graph-based memory systems. He also founded the Pals of Autonomous Agents community (1,500+ members), driving cross-org alignment on AI agent best practices and production-ready LLM-based systems.
Alex was the first data science hire at Uber's Advanced Technologies Group (self-driving cars), where he holds two patents and built safety-critical AI systems—from ML-powered semantic scene understanding combining NLP and computer vision across 1,000+ TB of sensor data, to map automation that saved $10M per year. He went on to ship enterprise AI products at SambaNova Systems and C3.ai, including SambaNova's premier large language model product and C3.ai's core data science offering for Bank of America, Air Force Research Lab, and Royal Dutch Shell.
- →Leading US Exploration team (12 people) for Seed's frontier AI models at ByteDance, from generation 1.6 to Seed 2.0
- →Incubated and open-sourced GraphRAG (32k stars), Semantic Kernel (27k stars), and AutoGen (54k stars) at Microsoft
- →First data science hire at Uber ATG — two patents, safety-critical AI for self-driving cars
- →Published: Graph RAG for Query-Focused Summarization; Landscape of Emerging AI Agent Architectures
- →UC Berkeley (Statistics & CS) · Harvard Business School (Strategy & Leadership)
After a decade building AI—from self-driving cars at Uber to open-source frameworks at Microsoft to frontier models at ByteDance—I noticed a pattern: the most impactful work happened when builders stayed connected to the people they were building for. The failures came when teams optimized for technical elegance over human outcomes.
I started Humans of AI to keep that human thread visible—through stories that reveal how practitioners actually work, tools that solve real problems, and investments in founders who refuse to lose sight of who they're serving.
"I've built AI for self-driving cars, open-sourced frameworks used by millions, and led frontier model research. Now I want to tell the stories of the people doing this work—and back the ones who are doing it right."
Help us grow
Humans of AI is an independent project. Here's how you can help us reach more people and build more thoughtfully.
Watch & Share
Subscribe to the playlist on YouTube and share your favorite episodes with someone who would appreciate them.
View Playlist →Nominate a Guest
Know someone whose AI story should be heard? Share the story, concrete work, and point of view that would make a strong episode.
Suggest a Guest →Sponsor the Show
We talk to the people actually building AI. Our audience is the one they're trying to hire, sell to, and partner with. If that's who you want to reach, let's talk.
Get in Touch →Collaborate
Want to build tools, create content, or explore new formats together? We're always open to thoughtful collaboration.
Get in Touch →Get notified about new episodes
We'll email you when new episodes drop, along with occasional updates on what we're building. No spam, just substance.
Subscribe on Substack →Unsubscribe anytime. We respect your inbox.