Edition #39: AI is leaving the screen. Who decides what it's for?

A former Google robotics leader on the societal storm ahead, why dignity is the real design goal, and the one exercise that makes everything simpler

By Mykel Salomon · 2026-09-25

<hr><h3>🔓 Opening Reflection</h3>Every weekend, <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 ember-view" href="https://www.linkedin.com/in/hanspeter/">Hans Peter Brøndmo</a> would call his mother in Norway from California. And every weekend, she would pick up the phone and skip the hello entirely."Hans Peter, where are the robots?"It became a ritual between them. He would say they're coming, Mom, but it's going to take a while. And she would answer: well, they better come soon.They didn't come soon enough. She passed away a few years ago, living on her own until the end, supported by three government-funded caregivers who came by every day: morning, midday, early evening. They helped her dress, bathe, and tidy up. But between those visits were long, quiet hours. If her book slipped to the floor, she couldn't pick it up. So she waited, sometimes three hours, unable to read, for someone to walk through the door.What she wanted from a robot wasn't novelty. It was dignity.And here is the part that stopped me. Hans Peter imagined what those visits could have been if a machine had handled the book, the dishes, the small tasks we all take for granted. The human who walked in could have sat down and played cards with her. Taken her for the short walk around the neighborhood she loved.That story is the whole edition. Because Hans Peter spent seven years leading Everyday Robots, Google X's moonshot to build AI-powered helper robots, and he believes AI is about to step off our screens and into the physical world: our hospitals, our homes, our institutions. The question he keeps asking isn't how smart the machines will get.It's who gets to decide what they're for.<hr><h3>🔦 Signal of the Week</h3>We keep calling this the future of work. Hans Peter thinks that framing is far too small.Look at the last time technology amplified human capability at this scale. The Industrial Revolution didn't just change factories. For roughly two generations, it was chaos. Children worked and died in coal mines. Cities swelled with no planning. And economic historians have a name for what happened to ordinary workers: Engels' pause, a stretch from about 1790 to 1840 when British output per worker climbed while working-class wages stagnated, and the profit rate doubled. Between 1780 and 1840, output per worker rose around 46%. Real wages rose about 12%.The gains eventually reached people. But only after society invented what it needed to share them: labor laws, unions, public institutions, new ideas about what a worker is owed.Now consider the scale of what's coming. <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 eBfrkeSTVtlOgnfXzKdhfMqooLsYbETjEnlcIg " href="https://www.linkedin.com/company/google/">Google</a> <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 eBfrkeSTVtlOgnfXzKdhfMqooLsYbETjEnlcIg " href="https://www.linkedin.com/company/googledeepmind/">Google DeepMind</a> 's <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 ember-view" href="https://www.linkedin.com/in/demishassabis/">Demis Hassabis</a> has said AI could be ten times bigger than the Industrial Revolution, arriving maybe ten times faster. And when that first revolution began, there were roughly one billion people on Earth. Today there are over eight billion, and this transition will hit all of them, unevenly, at different speeds, in different places.So when the public conversation shrinks AI down to safety legislation and job losses, Hans Peter holds up his fingers an inch apart. Those matter, he says. But they're this much of something enormous.We're not preparing for a new product cycle. We're preparing for a new society.<hr><h3>🌍 In the Wild</h3><h3>1. You can't program a robot to live among us. You have to teach it.</h3>This was the insight that shaped Hans Peter's years at Google X, back in early 2016, a full year before the Transformer paper that gave us today's language models.Look at a glass of water on a table. You can see through it. So how does a machine know it's a solid object it can pick up? Turn your face toward a window and the light changes everything a camera sees. Wait until evening and the whole room looks different. You cannot write enough rules to cover the real world. There are too many edge cases, too much ambiguity.The breakthrough came from translation. For decades, people tried to encode grammar as rules and it never quite worked. Then machine learning figured it out from data. The team's bet was simple: what if we could do the same for robots? Stop programming them. Teach them. (Early on, one of their robots learned to sort office trash, cutting recycling contamination from about 20% to under 5%.)That's why Hans Peter calls robots the physical manifestation of AI. The same intelligence we chat with is learning to act, which means the decisions we make about AI are no longer only about words on a screen. They're about bodies in the world.<h3>2. Plan for the storm you can't predict</h3>I asked him to describe the AI society we're heading toward. He refused to pretend he knew, and I respected him more for it. Anyone predicting the future today, he said, is working from personal belief, not knowledge. Tell any story you like, utopia or apocalypse, and you'll find a passionate expert who swears you're right.So he offered a metaphor that landed hard for someone like me, living in Miami.We're in the early stages of an AI hurricane. Something is forming off the coast. The models update every few hours. Nobody knows exactly where it makes landfall, or whether it arrives as a stiff breeze or a catastrophe. But nobody on the Gulf Coast waits for certainty before preparing. You know the season comes every year. You strengthen the infrastructure. You make a plan.That's where we are: the pre-planning stage. We can't predict the future. But we can decide, now, what we're building toward, and prepare for a series of very big storms. (His one caveat: unlike a hurricane, this storm can be harnessed for enormous good. He's an optimist. Just not a naïve one.)<h3>3. We need a New Deal for the AI society</h3>Hans Peter has been, in his words, geeking out on the New Deal. And the part most people miss is that it wasn't just a list of programs. It was built on humanist principles: that people have a right to exist outside of work, that a human being is not a unit of labor to be discarded once a company no longer needs them.The echoes to today are uncomfortable. FDR called the concentrated wealth of his era "economic royalists." We are living through our own debate about tech billionaires, inequality, and who captures the gains. And just like Engels' pause, the question isn't whether value gets created. It's whether it reaches people, and how long they have to wait.His argument is that someone needs to tell a story that acknowledges the hardship ahead, names it honestly, and has a plan for it. He even floated it as a political platform: a leader who says AI is a necessary accelerant for our economy and our future, and that it must be deployed on humanist principles. Safety rules and cross-border agreements are necessary. But they are a sliver of the real task, which is designing a society around people first.Right now, he says, no one is convening that conversation at the scale it deserves. That's exactly why it can't stay on podcasts.<h3>4. Fewer people means we need the machines</h3>Here is the part of the AI debate that almost nobody talks about, and it reframes everything. Hans Peter says AI and robotics won't just happen. They must happen.When a population's fertility rate falls below about 2.1 children per couple, the decline doesn't drift down gently. It compounds, because each generation has fewer parents. South Korea hit 0.72 in 2023, the lowest rate ever recorded anywhere, and even with a modest rebound to around 0.80, its population is projected to fall from about 51 million to 36 million by 2072. Japan now has roughly 30% of its population over 65, and faces a shortage of hundreds of thousands of care workers in the coming decades.Fewer young people means fewer hands to care for more elders, fewer workers to produce what society needs, more pressure on pensions and healthcare. So countries like Korea and Japan are racing toward automation not out of greed, but necessity.There was a telling detail here. China just enacted rules restricting AI companions and virtual romantic partners for minors and banning services that foster emotional dependence at the expense of real relationships. Hans Peter's read: it's partly a demographic policy. A country facing a steep population decline would rather its young people fall in love with each other than with a chatbot.The silver lining he sees: fewer people working could mean less strain on the planet, and more time for people to be people. In community. Together.<h3>5. The real design goal is dignity</h3>Which brings us back to his mother, and to the question I asked him directly: what does protecting human dignity look like when machines become our assistants, our teachers, our workers?His answer was the phone ritual. Where are the robots? Not a robot to replace her caregivers. A robot to handle the book on the floor and the dishes in the sink, so that the human who walked in could spend the visit on what only a human can offer: a game of cards, a walk, company.That isn't just a hopeful vision. There's new evidence behind it. A Stanford study published in August 2026 looked at care robots in Japanese nursing homes and found they helped ease care-labor shortages without displacing workers or lowering care quality, and in some cases improved staff retention. The machines took some of the load. The humans stayed.This is the version of AI worth fighting for. Not machines instead of people. Machines that free people to be more human with each other.<h3>6. Stop selling fear. Aim for remarkable.</h3>Hans Peter said something that I think every leader, journalist, and politician needs to hear. Fear is one of the oldest motivators we have. If a tiger was on the prowl, fear kept you alive. But that survival instinct is being weaponized, in the news and in politics, and statistics about progress don't stand a chance against it.He isn't asking anyone to ignore real concerns. People losing work is a real concern. But we are ingenious, he said. We put a person on the moon. We learned to pull energy from the sun into our kitchens. Let remarkable be the bar, not dread.And he practices what he preaches: studying the New Deal, he got from an AI in minutes what might have taken a PhD months to explain, while staying alert enough to catch it when it invents things. Dario Amodei's phrase, "a country of geniuses in a data center," is becoming real. The human job is to remain wise, critical, and curious enough to use it well.Lean into wonder. Just keep your eyes open.<hr><h3>💬 The Big Question</h3>When AI steps off the screen and into our homes, hospitals, and institutions, who gets to decide what it's for?And the deeper one, which Hans Peter keeps returning to:If we could write the values of an AI society on a single page, like a declaration of human rights, what would we put first?His answer is humanism. Put the human first, then make every choice explain how it serves that. Simple to say. Hard to live. Worth insisting on.<hr><h3>🧠 A Small Exercise</h3>This is the practice Hans Peter does twice a year, and it's the best thing I took from our conversation. Give it an hour.<ol><li>Write down every word that represents a value to you. Love, connection, health, income, adventure, kindness. Fill the page.</li><li>Pick your ten most important.</li><li>Stack rank those ten. First, second, third, all the way down.</li><li>Keep the top three to five somewhere you'll see them. A sticky note on the fridge. Your phone's lock screen.</li><li>Now the hard part: open your calendar. Map what you actually do against those values. Where something doesn't fit, ask why. Either you're not being honest about what you value, or you're making choices out of alignment with it.</li></ol>He says the surprise is that living this way makes life simpler, not harder. The donut, the speaking invitation, the extra meeting: each one becomes an easy question. Does this serve what I said matters?That's intentionality. And the smarter the machines get, the more it becomes the most human thing we have.<hr><h3>🎙️ This Week on The Human Protocol Podcast</h3>Where Are the Robots? Physical AI, the Coming Societal Shift, and Designing for Dignity Guest: <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 ember-view" href="https://www.linkedin.com/in/hanspeter/">Hans Peter Brøndmo</a> , former VP at Google and head of Everyday Robots, Google X's AI robotics moonshot; technology entrepreneur; author of the forthcoming book Being Human in an Age of Intelligent MachinesHans Peter spent seven years at the exact edge where AI stops being software and starts acting in the real world. He brings the rare combination of a builder's experience and a humanist's conscience.In this episode, you'll learn:<ul><li>Why you can't program a robot to live among us, and why you have to teach it</li><li>Why the AI transition is a societal shift bigger than the Industrial Revolution, not just a jobs story</li><li>The hurricane metaphor: how to prepare for a future no one can predict</li><li>Why we need a New Deal for the AI society built on humanist principles</li><li>How falling birth rates make robotics a necessity, not a luxury</li><li>His mother's question, "where are the robots?", and what it teaches about dignity</li><li>The values exercise he uses twice a year to live with intention</li></ul>🎧 Available now on <a target="_self" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 eBfrkeSTVtlOgnfXzKdhfMqooLsYbETjEnlcIg " href="https://open.spotify.com/episode/03MDpfTcXZPV5GQkal53fs?si=0FcZAJfOTpiOD7LgLTruYA">Spotify</a>, <a target="_self" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 eBfrkeSTVtlOgnfXzKdhfMqooLsYbETjEnlcIg " href="https://podcasts.apple.com/us/podcast/ai-is-about-to-leave-the-screen-who-gets-to-drive-it/id1843402023?i=1000791278037">Apple Podcasts</a>, and <a target="_self" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80 eBfrkeSTVtlOgnfXzKdhfMqooLsYbETjEnlcIg " href="https://youtu.be/dbNVPnIlcvY?si=HdT2tIsLZI5W5Imo">YouTube</a>.<img class="max-w-full h-auto rounded-md my-4" src="https://media.licdn.com/dms/image/v2/D4E12AQGniqtSXJNF3g/article-inline_image-shrink_1000_1488/B4EaDYW.ODJMAI-/0/1790336296119?e=1792022400&v=beta&t=MWPNWMjPuh7TQQHE8k5SywAE1Z2Mw7L0adXc5UqnRfQ" alt="Article content"><hr><h3>🧭 Thought to End the Week</h3>When I asked Hans Peter what he'd tell a younger version of himself, he didn't talk about coding or engineering degrees. He said education is a time to learn how to learn, how to think, and how to be human. Study your field, yes. But also keep an art practice. Read philosophy. Take dance lessons. Be in your body. Be connected to the long, rich history of what humanity has figured out.Because one day, he said, you'll have power. You'll make decisions that matter for other people. And grounding those decisions in more than the science, in a deeper connection to people and to human wisdom, matters more now than ever.That's the whole thing, isn't it? The robots are coming. Some will look like us, most won't. They'll pick up the book from the floor, sort the trash, carry the load. The technology will keep getting better, faster than any of us can predict.What we decide it's for is still up to us.A woman in Norway already told us the answer. She didn't want a machine to replace the people who cared for her. She wanted a machine that would give those people back to her.Build for that.The code may change. And it will.But our humanity is the constant.See you next week.Mykel<hr><h3>📚 Sources</h3><ol><li>Hans Peter Brondmo, former VP at X (formerly Google X) and general manager of the Everyday Robots project from 2016; background and the "teach, don't program" approach. <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80" href="https://en.wikipedia.org/wiki/Hans_Peter_Brondmo">https://en.wikipedia.org/wiki/Hans_Peter_Brondmo</a> and <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80" href="https://techcrunch.com/?p=1915354">https://techcrunch.com/?p=1915354</a></li><li>Robert C. Allen, "Engels' pause: Technical change, capital accumulation, and inequality in the British industrial revolution," Explorations in Economic History (2009): real wages stagnated from roughly 1790 to 1840 while output per worker grew and the profit rate doubled. <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80" href="https://www.sciencedirect.com/science/article/abs/pii/S0014498309000199">https://www.sciencedirect.com/science/article/abs/pii/S0014498309000199</a></li><li>South Korea fertility: record low of 0.72 in 2023, rebounding to about 0.80 in 2025; population projected to fall to 36.2 million by 2072. <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80" href="https://bworldonline.com/world/2026/02/25/732803/south-koreas-birthrate-the-worlds-lowest-rises-again-amid-signs-of-easing-demographic-crisis/">https://bworldonline.com/world/2026/02/25/732803/south-koreas-birthrate-the-worlds-lowest-rises-again-amid-signs-of-easing-demographic-crisis/</a></li><li>Stanford FSI, "Robots Ease Elder Care Shortages Without Displacing Workers or Diminishing Care Quality" (August 2026): evidence from Japanese nursing homes, where about 30% of the population is 65 or older. <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80" href="https://fsi.stanford.edu/news/robots-ease-elder-care-shortages-without-displacing-workers-or-diminishing-care-quality-study">https://fsi.stanford.edu/news/robots-ease-elder-care-shortages-without-displacing-workers-or-diminishing-care-quality-study</a></li><li>China's Interim Measures for AI Anthropomorphic Interactive Services (effective July 15, 2026): restrictions on virtual companion and partner services for minors and on fostering emotional dependence. <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80" href="https://www.cnn.com/2026/07/23/business/china-ai-companion-ban-intl-hnk">https://www.cnn.com/2026/07/23/business/china-ai-companion-ban-intl-hnk</a></li><li>Franklin D. Roosevelt, 1936 Democratic National Convention acceptance speech ("economic royalists"), and the New Deal's humanist foundations.</li><li>Dario Amodei, Machines of Loving Grace (2024), source of "a country of geniuses in a datacenter." <a target="_blank" rel="noopener noreferrer nofollow" class="text-[#4FD1C5] underline hover:opacity-80" href="https://www.darioamodei.com/essay/machines-of-loving-grace">https://www.darioamodei.com/essay/machines-of-loving-grace</a></li></ol>