The Technically Human Podcast

What happens to mathematics when doing mathematics is no longer an exclusively human activity?

For years, we've watched artificial intelligence get progressively better at math: solving equations, passing exams, succeeding at mathematical competitions, and generating proofs. But recent developments seem to cross a different kind of threshold. This month, OpenAI announced that an internal AI system had produced a solution to the Navier–Stokes existence and smoothness problem—one of the seven Millennium Prize Problems, designed to represent some of the most difficult unsolved questions in mathematics.

The announcement has generated excitement, skepticism, and controversy. But it also raises a much larger set of questions. What does it actually mean to do mathematics? Is mathematics about arriving at a correct answer, or understanding why that answer is true? What happens to ideas like discovery, creativity, and authorship when machines can participate in producing new mathematical knowledge? And what does that tell us about artificial intelligence—and about ourselves?

My guest today is Dr. Noah Giansiracusa, a math professor at Bentley University and a faculty associate at Harvard, with a PhD in algebraic geometry from Brown University.

Noah is cohost of the podcasts Breaking Math and AI in Academia, and his writing has appeared in The Washington Post, Scientific American, TIME, Wired, Slate, and The Boston Globe, among others.

His first book, How Algorithms Create and Prevent Fake News, examined the mathematics and technology underlying our information ecosystem. His latest, Robin Hood Math, asks how people from all backgrounds can use mathematics to navigate—and reclaim some control over—a world increasingly governed by algorithms.

Today, we're talking about what happens when those algorithms begin doing mathematics themselves.

What does it mean for the web to be open—and open to whom?

For most of the web’s history, we’ve operated on a fairly simple assumption: if you put something on a public website, people could find it, search engines could index it, and information could move relatively freely across the internet. But as AI systems increasingly crawl, extract, learn from, and act on information from the web, that bargain is beginning to change.

My guest today is Shane Evans, co-founder and CEO of Zyte, a company that builds technology for extracting data from the public web for use in AI, analytics, and other applications.

Shane has been working on the web since the late 1990s. He was a lead engineer at lastminute.com during the dot-com era, later created Scrapy, the widely used open-source web-scraping framework, and founded Zyte in 2010.

Today, we’re talking about Zyte’s new research suggesting that the Open Web is increasingly becoming a “Controlled Web”—and what that means for publishers, AI companies, web users, and the future of access to information online.

My guest today is Dave Merritt, cofounder and Chief Technology Officer of Watch Duty.

Founded in 2021, Watch Duty is a nonprofit emergency-information platform that provides real-time, verified intelligence about wildfires and floods across all 50 states. Its network of active and retired firefighters, dispatchers, first responders, and reporters monitors radio traffic, satellites, cameras, weather data, and official sources to give people information they can act on—often in the critical period before traditional emergency communications catch up. Today, Watch Duty serves more than 16 million active users annually.

Building technology people rely on during a disaster raises profound questions about trust: How do you move information quickly without sacrificing accuracy? How much uncertainty should you expose to the public? What should we automate—and what decisions still require human judgment? And what happens when a nonprofit technology platform begins functioning as critical public infrastructure?

In our conversation, we talk about what technological infrastructure looks like when wildfire risk has increasingly become part of the background condition of everyday life. When a fire does start, one of the most consequential questions is also one of the simplest: What is actually happening, and what do I need to do?

Watch Duty was built to answer that question.

In the wake of this week’s landmark $18 billion settlement with Meta, we’re bringing you an episode with Yaël Eisenstat, one of the leading thinkers on the issues at the heart of that case — and the many other recent legal challenges that may be fundamentally changing how we think about the power and responsibility of technology platforms.

In recent months, courts and juries across the country have been confronting questions that technology policy has wrestled with for years: When should platforms be held responsible for harms facilitated by their products? Where is the line between regulating speech and regulating product design? And what does it look like to build the legal cases that hold Big Tech accountable?

And after decades in which Section 230 and other legal protections often prevented these questions from reaching juries at all, are we witnessing the beginning of what we’re calling tech’s Big Tobacco moment?

In the episode, I speak with Yaël Eisenstat about the recent cases against Meta and other platforms, the years of research, journalism, whistleblowing, advocacy, and litigation that made this moment possible, and what happens when evidence about the consequences of technological design finally reaches ordinary jurors.

We also ask what these cases can—and cannot—accomplish. Can litigation succeed where legislation and regulation have struggled? Can social media actually be redesigned to serve the public interest, or are its harms inseparable from its underlying business model? And as our technological landscape shifts from social media toward increasingly personalized and intimate AI systems, are we building an accountability framework for the last technological era just as the next one arrives?

Yaël Eisenstat is the Director of Policy and Impact at the CyberSafety Research Center, where she works on policy approaches to holding online platforms accountable for their effects on public safety and democracy. She previously served as Vice President of the Anti-Defamation League’s Center for Technology & Society and, in 2018, joined Facebook as its Global Head of Elections Integrity Operations for political advertising. Before entering technology policy, Eisenstat spent 14 years in public service as an intelligence officer, diplomat, and White House adviser. She has since become a prominent advocate for greater transparency and accountability in the technology industry.

My conversation today is with Monika Bielskyte, a futurist, foresight strategist, speculative designer and founder of Protopia Futures. She developed the Protopia World Stack, a science-informed, life-centric foresight methodology. Protopia is an Inclusive future vision based on evolved values, informed by bleeding edge scientific research and supported by advanced technological tools.

She was Nike's first Futurist in Residence, is an EY Luminary, and has consulted across entertainment, technology, science, and government – including Disney and Marvel's Black Panther franchise. Her technical expertise spans AI, spatial computing, XR, and synthetic media, with particular attention to how these systems can expand human agency or be weaponized for behavioral manipulation and disinformation.

Monika bridges the worlds of entertainment and real future-making by prototyping culturally expansive, socially and environmentally engaged future visions that combine the best of creativity and science, delivering energizing vision, lasting value and resilience to disruptions. What connects all of it is a commitment to futures literacy - the belief that rendering better futures compelling, curiosity-inducing, and inhabitable is the mechanism by which transformation happens.

On this episode I’m joined by Jess Holbrook, Head of UX Research for Microsoft AI.

We talk about what UX research can—and cannot—tell us about what people need; why user-centered design may be too individualistic for technologies whose effects extend across families, workplaces, communities, and institutions; and what it means to put researchers in the room while AI systems are still being built rather than after fundamental design decisions have already been made.

Over the past 12 years, Jess has held senior research and design leadership roles across Microsoft, Meta, Google, and Amazon. He previously led Generative AI, Responsible AI, and data transparency UX research at Meta, and at Google he founded and led the People + AI Research group, or PAIR, before serving as Head of UX for the Responsible AI and Human-Centered Technology group. His work has focused on bringing human-centered and inclusive approaches into the development of AI systems and translating emerging machine-learning capabilities into products designed around people’s needs.

Jess also writes regularly about AI, frontier UX research, research slop and the future of human-centered technology on his Substack, Gen Purpose.

Today I’m speaking with Dana Mazia, General Manager of the Bright Initiative by Bright Data, which provides nonprofits, academic institutions, and public bodies with pro bono access to public web data and expertise.

Dana is a lawyer, social entrepreneur, and World Economic Forum Global Shaper with more than 14 years of experience across technology, data, and social impact.

Bright Data is a web-data platform that provides the infrastructure organizations use to collect, process, and analyze publicly available information from across the internet. The Bright Initiative is its public-interest program, making that technology and expertise available to organizations working in academic research, public policy, transparency, and social impact.

Conversations about artificial intelligence typically focus on models: what they can do, what they produce, and whether their outputs are accurate or safe. But behind every model is a less visible infrastructure—the systems through which information is collected, organized, and made usable.

That raises important questions about who can access public web data, what “public” really means, and whether an open internet benefits everyone equally or primarily its most powerful actors. These questions become even more urgent as AI systems move from learning from the web to actively operating within it.

Dana and I discuss the politics of data infrastructure, the ethics of web scraping, the risk of information monopolies, and the legal and ethical questions that arise when information from the public web is collected and operationalized at scale. Ultimately, we ask who should have access to this infrastructure—and who will have the power to shape the AI ecosystem built upon it.

Today I am interviewing Bruce Friedrich, the author of the newly published book "Meat: How the Next Agricultural Revolution Will Transform Humanity's Favorite Food―and Our Future." Bruce Friedrich is the president and founder of the Good Food Institute, one of the leading think tanks working to make the global food system better for the planet, people, and animals. As an international network of organizations, they advance alternative proteins to help meet climate, global health, food security, and biodiversity goals.

My guest for this episode is Professor Aleem Hossain, a writer and director working across science fiction, fantasy, experimental cinema, and television. His feature film After We Leave won Best Feature Film at the Sci-Fi-London film festival and Best Director at the Berlin Sci-Fi festival, and his AI-generated short film, Do Bangladroids Dream of Electric Tagore?, premiered at Fantastic Fest and won Best Narrative Film at the AI Art & Film Awards.

His work is shaped by his own multicultural upbringing—the son of a Muslim immigrant from Bangladesh and a mother from a Roman Catholic family in Connecticut—as well as by his interest in expanding both the kinds of stories we tell and the people who are able to tell them.

He is also an associate professor of digital storytelling at Occidental College, where he teaches filmmaking and screenwriting.

In this conversation, we talk about what generative AI makes possible for independent filmmakers, what it fails to understand about culture and identity, and how the arrival of these tools may change not only how films are made, but who gets to imagine the future.

This episode was produced by Jack Lucas Chang.

Globally, 8 billion people will experience trauma in their lifetime. Ninety-three percent of those in crisis will not receive adequate care. The barriers are familiar: cost, stigma, language, geography, workforce shortages. What is less familiar is someone actually building technology to address all of them at once, starting from survivor experience rather than market opportunity.

Megs Shah is the Founder and CEO of The Parasol Cooperative, a nonprofit that builds trauma-informed AI tools for survivors of domestic violence, human trafficking, and technology-facilitated abuse. Her flagship product, Ruth, is the world's first trauma-informed AI chat — designed by and for survivors and the advocates who support them. In just over a year, Ruth has facilitated more than 679,000 private crisis support interactions across 147 countries in 85 languages, earning a 92% helpfulness rating. It requires no login, no account, and no personal information. Privacy is not a feature; it is the architecture.

In this conversation, Megs and I talk about what it means to build technology under a fundamentally different moral and financial logic — one that centers dignity, safety, and access rather than engagement and data capture. We discuss how lived experience from survivors shapes design decisions that conventional technologists miss, why trauma-informed AI is categorically different from general-purpose AI deployed in sensitive contexts, and what it looks like to measure success when the meaningful metrics are trust, stabilization, and safety rather than clicks. We also talk about the ways technology is already being weaponized against the people Ruth is built to protect — and what it would take to design systems that take patterned harm seriously before it escalates.

This episode was produced by Jack Lucas Chang.