The Gap and You: AI, Personal Context, and Aristotle
The Gap and You
AI, Personal Context, and Aristotle
By Yubei Chen | Aizip
In 1086, twenty years after the Norman Conquest, William I’s commissioners traveled across England asking questions. Who owned the land? What was on it? What was it worth? Their answers became Domesday Book, a detailed record of the country’s wealth and ownership. [1]
In A History of Britain, Simon Schama described the effect:
“as if he’d reconquered England—statistically. Because its information was more impregnable than any castle.” [2]
Nearly a millennium ago, information was already an instrument of power.
Today, AI brings that relationship into our daily lives. An assistant becomes more useful when it understands not just what we ask, but why we are asking. That background—our experiences, relationships, responsibilities, and goals—is our personal context.
Who controls it, and who benefits from it?
One person, two forms of payment
Centuries later, in Edo—the city now called Tokyo—Kasai Gonshirō brought two boats to the castle each morning to collect an unlikely cargo: human waste. He took it back to his village, where it could fertilize the fields. What the city needed removed, farmers wanted. [3]
His boats connected two needs. It was an unglamorous example of a familiar business idea: make an exchange easier and more valuable for everyone involved.
Some successful businesses take this a step further: create enough value for both sides that both are willing to pay.
The pattern appears across eras. Nineteenth-century newspapers sold copies to readers and advertising space to businesses. [4] Auction houses collected fees from buyers and sellers. [5] Online accommodation marketplaces later charged both hosts and guests. [6] Different businesses, but the same basic arrangement: serve two sides and earn revenue from both.
The digital era added another possibility: the same person could pay twice—first with money, and then, with data.
Alongside a purchase or subscription, we supply information about our habits, interests, and relationships. That information can help a platform sell advertising and improve its services. It does not appear on the invoice, but it has value. [7]
Much of this context has been scattered: purchases in one place, relationships in another, documents and messages somewhere else. As platforms connect services and combine information, those pieces can form a more complete picture of the same person. [7]
AI gives us an even stronger reason to bring those pieces together.
We pay for intelligence while supplying the context that makes it useful. The more we want an assistant to handle, the more we have a reason to share.
We are no longer contributing only records of what we do. We are explaining why we do it, what matters to us, and how we make decisions. Over time, information scattered across our lives can become a detailed understanding held by one service.
That can make assistance much better. But getting better help should not require giving up control of what makes that help possible.
The last few tokens
Consider an experienced account manager answering a long-standing customer.
An AI drafts a polished response. The manager corrects it: don’t promise Friday until the warehouse confirms. Acknowledge the late delivery last time. This customer needs a phone call, not another reassuring email.
Think of these as “the last few tokens”—the small but crucial pieces of context and judgment an AI still needs from us to finish the job.
They are not a fixed boundary. What requires a person today may become an automated step tomorrow.
Each correction makes some expertise easier to describe. When those examples become training material, they can help AI reproduce work it previously could not do. Research on customer-support AI already suggests that systems can spread experienced workers’ practices to less-experienced colleagues. [8]
We may be paying for assistance while paying with data to train systems to do more of our work without us. The second payment could be more expensive.
The employment consequences are not merely hypothetical. Through September 2026, U.S. employers cited AI in more than 120,000 announced job cuts, according to Challenger, Gray & Christmas. [9]
A job does not have to be fully automated for someone to lose it. If a smaller team can deliver the same work, positions can disappear while some tasks still require human judgment.
The people whose knowledge helped create those capabilities may have little say in how they are used—or share in the returns.
We risk becoming more dependent on systems that are becoming less dependent on us.
The Matrix offers a useful metaphor here—not a prediction. Are we building intelligence that serves us, or turning our work, decisions, and memories into raw material for systems we do not control?
What makes you you?
Think of a lesson you learned the hard way. A connection you noticed because of an unusual experience. An idea you have not yet found the words to explain.
Your history shapes what you see and what you can bring into the world.
We believe there is original value in every person that the rest of the world cannot simply reconstruct without their participation. We are not just records of what has already happened. We remain sources of what can happen next.
AI may learn to reproduce parts of our knowledge and judgment. That makes control over our context more important, not less. As we explain what we know and how we think, we can make some of our contributions reproducible without our continued involvement.
Who then controls those capabilities? Who benefits from them?
The concern is not that sharing knowledge diminishes our worth. It is that we could help build extraordinary intelligence while losing control over the value our lives helped create.
Protecting that value does not mean keeping knowledge locked away. It means giving people tools to develop their ideas, choose how to share them, and benefit from what they make possible.
Our context should help us create new possibilities—not simply make our contribution easier to replace.
A revolution people can participate in
We are in the middle of a technological revolution. At Aizip, we want to help it move forward: removing tedious work, making expertise more accessible, and helping people accomplish things previously beyond their reach.
Protecting people’s value does not mean preserving every task as a human job. It means helping people use better tools to create, own, and benefit from new value.
Even if AI brings an age of abundance, cheaper goods and services would not automatically give everyone ownership or independence. Our children should have meaningful ways to build and participate—not simply subscriptions to services whose terms they cannot influence.
An optimistic future should not depend entirely on the goodwill of a few platform owners.
The goal is not to keep people indispensable to every task. It is to keep the future in people’s hands.
The unfinished promise of personal computing
In June 1983, Steve Jobs stood before an audience of designers in Aspen and described what books had meant to him as a student. He could read Aristotle and Plato directly, reaching across centuries to encounter their ideas.
But there was a problem: “you can’t ask Aristotle a question.”
Jobs asked his audience to imagine the next Aristotle carrying a computer through life. As this person recorded thoughts and ideas, perhaps the machine could capture something of their way of seeing the world. Long after they were gone, someone might ask: “What would Aristotle have said?” [10]
The possibility was not just preserving someone’s words, but continuing a conversation.
This vision sat alongside another ambition Jobs had expressed: computers that ordinary people could buy, explore, and understand for themselves. [10]
For us, that promise has an unfinished part. A machine that understands you is not necessarily a machine that belongs to you.
Even owning the device is not enough if the knowledge and memories built through years of use remain locked into someone else’s service.
At Aizip, our part of this mission is to put more intelligence into everyday devices, close to people’s lives and under their control.
That is a difficult engineering challenge. Devices have limits in memory, power, heat, and cost. People want useful assistance, not a worse experience offered in the name of privacy. Local intelligence has to work well enough to earn its place.
Local-first does not mean cloud-never. We want to help build systems that keep personal memory and much of the processing on the user’s devices, while calling on outside services for tasks that need them—with the user choosing what information to share.
Keeping information nearby is only part of the answer. People should be able to see what an assistant remembers, correct it, delete it, and take it with them. Changing AI providers should not mean starting their digital lives over. Those expectations should apply to what we build, too.
Local intelligence will not solve every economic consequence of AI. But it is one practical way to give people more control over the knowledge and capabilities they help create.
We are working to make that option increasingly useful—in people’s pockets and in their hands. This mission is much larger than Aizip, and we welcome others building toward it.
Leave room for what stays with us
The poet and translator Anne Carson, winner of the 2026 Nobel Prize in Literature, offers a different way to think about what is missing. [11] In If Not, Winter, her translation of the surviving fragments of Sappho’s poetry, brackets and blank space mark where words have been lost. She lets the gaps remain visible. [12]
For us, those pages offer a reminder: not every gap needs to be filled.
In The Secret Life of Walter Mitty, photographer Sean O’Connell finally has a snow leopard before his camera. Walter asks when he is going to take the picture.
“Sometimes I don’t,” Sean says. “I just want to stay in it.” [13]
He is not rejecting photography. He is choosing when the camera should—and should not—be part of the experience.
Personal AI should preserve that choice.
Sometimes that means processing information and keeping memories on our own devices. Sometimes it means sharing selected context to get better help. Sometimes it means choosing not to record an experience at all.
The gap worth preserving is not a permanent limit on what AI can do. It is room for each person to bring something original into the world—and retain a say in what becomes of it.
We want AI to help people do more, understand more, and remember what matters, while keeping the decisions about their lives in their hands.
How can we let your memory stay here—with you?
References
[1] Hull Domesday Project. Domesday Book. Historical background on the survey, its scope, and the information collected. https://www.domesdaybook.net/ Domesday Book
[2] Simon Schama. A History of Britain, Series 1, Episode 2: “Conquest.” BBC. Quotation at approximately 52:30–52:38; transcript hosted by Subsaga. https://subsaga.com/bbc/documentaries/history/a-history-of-britain-by-simon-schama/series-1/2-conquest.html Subsaga
[3] Kayo Tajima. The Marketing of Urban Human Waste in the Edo/Tokyo Metropolitan Area: 1600–1935. PhD dissertation, Tufts University, 2005, p. 71. Includes the account of Kasai Gonshirō transporting waste from Edo Castle. https://www.researchgate.net/publication/272791701_The_marketing_of_urban_human_waste_in_the_EdoTokyo_metropolitan_area_1600-1935 ResearchGate
[4] University of Illinois Library. American Newspapers, 1800–1860: City Newspapers. Discusses newspaper sales, circulation, and advertising revenue. https://www.library.illinois.edu/hpnl/tutorials/antebellum-newspapers-city/ University Library
[5] Christie’s. Understanding Auction Fees. Explains buyers’ premiums and sellers’ commissions. https://www.christies.com/en/help/auction-help-library/understanding-auction-fees Christie's
[6] Airbnb Help Center. Airbnb Service Fees. Describes the split-fee structure charging both hosts and guests, alongside the transition to a single-fee structure. https://www.airbnb.com/help/article/1857 Airbnb
[7] Federal Trade Commission. FTC Staff Report Finds Large Social Media and Video Streaming Companies Have Engaged in Vast Surveillance of Users with Lax Privacy Controls and Inadequate Safeguards for Kids and Teens. September 19, 2024. https://www.ftc.gov/news-events/news/press-releases/2024/09/ftc-staff-report-finds-large-social-media-video-streaming-companies-have-engaged-vast-surveillance Federal Trade Commission
[8] Erik Brynjolfsson, Danielle Li, and Lindsey Raymond. Generative AI at Work. arXiv:2304.11771, 2023; revised 2024. https://arxiv.org/abs/2304.11771 arXiv
[9] Challenger, Gray & Christmas. Job Cuts Fall in September; Hiring Plans Up 3% Over 2025 on Weak Early Seasonal Hiring. October 1, 2026. Reports 120,136 announced job cuts attributed by employers to AI through September 2026. https://www.challengergray.com/blog/job-cuts-fall-in-september-hiring-plans-up-3-over-2025-on-weak-early-seasonal-hiring/ Challenger Gray Christmas
[10] Steve Jobs. Make Something Wonderful: Steve Jobs in His Own Words. Edited by Leslie Berlin, Steve Jobs Archive, 2023. See “Speech at the International Design Conference in Aspen” (1983) and “Interview with The New Yorker” (1977). https://book.stevejobsarchive.com/ Steve Jobs Archive
[11] Associated Press. Canadian Author Anne Carson Wins Literature Nobel for “Bold and Inventive” Works That Blend Genres. October 8, 2026. https://apnews.com/article/05a268ffc76827d273f9d986c5a5a65e AP News
[12] Sappho. If Not, Winter: Fragments of Sappho. Translated by Anne Carson. Alfred A. Knopf, 2002; Vintage paperback, 2003. See also Dimitrios Yatromanolakis, “Fragments, Brackets, and Poetics: On Anne Carson’s If Not, Winter,” International Journal of the Classical Tradition 11, 266–272 (2004). https://www.penguinrandomhouse.com/books/160692/if-not-winter-by-sappho-translated-by-anne-carson/ https://doi.org/10.1007/BF02720036 Penguin Random House
[13] The Secret Life of Walter Mitty. Directed by Ben Stiller, 2013. Sean O’Connell’s dialogue in the snow leopard scene; quotation listed on IMDb. https://www.imdb.com/title/tt0359950/quotes/ IMDb
The Gap and You
AI, Personal Context, and Aristotle
By Yubei Chen | Aizip
In 1086, twenty years after the Norman Conquest, William I’s commissioners traveled across England asking questions. Who owned the land? What was on it? What was it worth? Their answers became Domesday Book, a detailed record of the country’s wealth and ownership. [1]
In A History of Britain, Simon Schama described the effect:
“as if he’d reconquered England—statistically. Because its information was more impregnable than any castle.” [2]
Nearly a millennium ago, information was already an instrument of power.
Today, AI brings that relationship into our daily lives. An assistant becomes more useful when it understands not just what we ask, but why we are asking. That background—our experiences, relationships, responsibilities, and goals—is our personal context.
Who controls it, and who benefits from it?
One person, two forms of payment
Centuries later, in Edo—the city now called Tokyo—Kasai Gonshirō brought two boats to the castle each morning to collect an unlikely cargo: human waste. He took it back to his village, where it could fertilize the fields. What the city needed removed, farmers wanted. [3]
His boats connected two needs. It was an unglamorous example of a familiar business idea: make an exchange easier and more valuable for everyone involved.
Some successful businesses take this a step further: create enough value for both sides that both are willing to pay.
The pattern appears across eras. Nineteenth-century newspapers sold copies to readers and advertising space to businesses. [4] Auction houses collected fees from buyers and sellers. [5] Online accommodation marketplaces later charged both hosts and guests. [6] Different businesses, but the same basic arrangement: serve two sides and earn revenue from both.
The digital era added another possibility: the same person could pay twice—first with money, and then, with data.
Alongside a purchase or subscription, we supply information about our habits, interests, and relationships. That information can help a platform sell advertising and improve its services. It does not appear on the invoice, but it has value. [7]
Much of this context has been scattered: purchases in one place, relationships in another, documents and messages somewhere else. As platforms connect services and combine information, those pieces can form a more complete picture of the same person. [7]
AI gives us an even stronger reason to bring those pieces together.
We pay for intelligence while supplying the context that makes it useful. The more we want an assistant to handle, the more we have a reason to share.
We are no longer contributing only records of what we do. We are explaining why we do it, what matters to us, and how we make decisions. Over time, information scattered across our lives can become a detailed understanding held by one service.
That can make assistance much better. But getting better help should not require giving up control of what makes that help possible.
The last few tokens
Consider an experienced account manager answering a long-standing customer.
An AI drafts a polished response. The manager corrects it: don’t promise Friday until the warehouse confirms. Acknowledge the late delivery last time. This customer needs a phone call, not another reassuring email.
Think of these as “the last few tokens”—the small but crucial pieces of context and judgment an AI still needs from us to finish the job.
They are not a fixed boundary. What requires a person today may become an automated step tomorrow.
Each correction makes some expertise easier to describe. When those examples become training material, they can help AI reproduce work it previously could not do. Research on customer-support AI already suggests that systems can spread experienced workers’ practices to less-experienced colleagues. [8]
We may be paying for assistance while paying with data to train systems to do more of our work without us. The second payment could be more expensive.
The employment consequences are not merely hypothetical. Through September 2026, U.S. employers cited AI in more than 120,000 announced job cuts, according to Challenger, Gray & Christmas. [9]
A job does not have to be fully automated for someone to lose it. If a smaller team can deliver the same work, positions can disappear while some tasks still require human judgment.
The people whose knowledge helped create those capabilities may have little say in how they are used—or share in the returns.
We risk becoming more dependent on systems that are becoming less dependent on us.
The Matrix offers a useful metaphor here—not a prediction. Are we building intelligence that serves us, or turning our work, decisions, and memories into raw material for systems we do not control?
What makes you you?
Think of a lesson you learned the hard way. A connection you noticed because of an unusual experience. An idea you have not yet found the words to explain.
Your history shapes what you see and what you can bring into the world.
We believe there is original value in every person that the rest of the world cannot simply reconstruct without their participation. We are not just records of what has already happened. We remain sources of what can happen next.
AI may learn to reproduce parts of our knowledge and judgment. That makes control over our context more important, not less. As we explain what we know and how we think, we can make some of our contributions reproducible without our continued involvement.
Who then controls those capabilities? Who benefits from them?
The concern is not that sharing knowledge diminishes our worth. It is that we could help build extraordinary intelligence while losing control over the value our lives helped create.
Protecting that value does not mean keeping knowledge locked away. It means giving people tools to develop their ideas, choose how to share them, and benefit from what they make possible.
Our context should help us create new possibilities—not simply make our contribution easier to replace.
A revolution people can participate in
We are in the middle of a technological revolution. At Aizip, we want to help it move forward: removing tedious work, making expertise more accessible, and helping people accomplish things previously beyond their reach.
Protecting people’s value does not mean preserving every task as a human job. It means helping people use better tools to create, own, and benefit from new value.
Even if AI brings an age of abundance, cheaper goods and services would not automatically give everyone ownership or independence. Our children should have meaningful ways to build and participate—not simply subscriptions to services whose terms they cannot influence.
An optimistic future should not depend entirely on the goodwill of a few platform owners.
The goal is not to keep people indispensable to every task. It is to keep the future in people’s hands.
The unfinished promise of personal computing
In June 1983, Steve Jobs stood before an audience of designers in Aspen and described what books had meant to him as a student. He could read Aristotle and Plato directly, reaching across centuries to encounter their ideas.
But there was a problem: “you can’t ask Aristotle a question.”
Jobs asked his audience to imagine the next Aristotle carrying a computer through life. As this person recorded thoughts and ideas, perhaps the machine could capture something of their way of seeing the world. Long after they were gone, someone might ask: “What would Aristotle have said?” [10]
The possibility was not just preserving someone’s words, but continuing a conversation.
This vision sat alongside another ambition Jobs had expressed: computers that ordinary people could buy, explore, and understand for themselves. [10]
For us, that promise has an unfinished part. A machine that understands you is not necessarily a machine that belongs to you.
Even owning the device is not enough if the knowledge and memories built through years of use remain locked into someone else’s service.
At Aizip, our part of this mission is to put more intelligence into everyday devices, close to people’s lives and under their control.
That is a difficult engineering challenge. Devices have limits in memory, power, heat, and cost. People want useful assistance, not a worse experience offered in the name of privacy. Local intelligence has to work well enough to earn its place.
Local-first does not mean cloud-never. We want to help build systems that keep personal memory and much of the processing on the user’s devices, while calling on outside services for tasks that need them—with the user choosing what information to share.
Keeping information nearby is only part of the answer. People should be able to see what an assistant remembers, correct it, delete it, and take it with them. Changing AI providers should not mean starting their digital lives over. Those expectations should apply to what we build, too.
Local intelligence will not solve every economic consequence of AI. But it is one practical way to give people more control over the knowledge and capabilities they help create.
We are working to make that option increasingly useful—in people’s pockets and in their hands. This mission is much larger than Aizip, and we welcome others building toward it.
Leave room for what stays with us
The poet and translator Anne Carson, winner of the 2026 Nobel Prize in Literature, offers a different way to think about what is missing. [11] In If Not, Winter, her translation of the surviving fragments of Sappho’s poetry, brackets and blank space mark where words have been lost. She lets the gaps remain visible. [12]
For us, those pages offer a reminder: not every gap needs to be filled.
In The Secret Life of Walter Mitty, photographer Sean O’Connell finally has a snow leopard before his camera. Walter asks when he is going to take the picture.
“Sometimes I don’t,” Sean says. “I just want to stay in it.” [13]
He is not rejecting photography. He is choosing when the camera should—and should not—be part of the experience.
Personal AI should preserve that choice.
Sometimes that means processing information and keeping memories on our own devices. Sometimes it means sharing selected context to get better help. Sometimes it means choosing not to record an experience at all.
The gap worth preserving is not a permanent limit on what AI can do. It is room for each person to bring something original into the world—and retain a say in what becomes of it.
We want AI to help people do more, understand more, and remember what matters, while keeping the decisions about their lives in their hands.
How can we let your memory stay here—with you?
References
[1] Hull Domesday Project. Domesday Book. Historical background on the survey, its scope, and the information collected. https://www.domesdaybook.net/ Domesday Book
[2] Simon Schama. A History of Britain, Series 1, Episode 2: “Conquest.” BBC. Quotation at approximately 52:30–52:38; transcript hosted by Subsaga. https://subsaga.com/bbc/documentaries/history/a-history-of-britain-by-simon-schama/series-1/2-conquest.html Subsaga
[3] Kayo Tajima. The Marketing of Urban Human Waste in the Edo/Tokyo Metropolitan Area: 1600–1935. PhD dissertation, Tufts University, 2005, p. 71. Includes the account of Kasai Gonshirō transporting waste from Edo Castle. https://www.researchgate.net/publication/272791701_The_marketing_of_urban_human_waste_in_the_EdoTokyo_metropolitan_area_1600-1935 ResearchGate
[4] University of Illinois Library. American Newspapers, 1800–1860: City Newspapers. Discusses newspaper sales, circulation, and advertising revenue. https://www.library.illinois.edu/hpnl/tutorials/antebellum-newspapers-city/ University Library
[5] Christie’s. Understanding Auction Fees. Explains buyers’ premiums and sellers’ commissions. https://www.christies.com/en/help/auction-help-library/understanding-auction-fees Christie's
[6] Airbnb Help Center. Airbnb Service Fees. Describes the split-fee structure charging both hosts and guests, alongside the transition to a single-fee structure. https://www.airbnb.com/help/article/1857 Airbnb
[7] Federal Trade Commission. FTC Staff Report Finds Large Social Media and Video Streaming Companies Have Engaged in Vast Surveillance of Users with Lax Privacy Controls and Inadequate Safeguards for Kids and Teens. September 19, 2024. https://www.ftc.gov/news-events/news/press-releases/2024/09/ftc-staff-report-finds-large-social-media-video-streaming-companies-have-engaged-vast-surveillance Federal Trade Commission
[8] Erik Brynjolfsson, Danielle Li, and Lindsey Raymond. Generative AI at Work. arXiv:2304.11771, 2023; revised 2024. https://arxiv.org/abs/2304.11771 arXiv
[9] Challenger, Gray & Christmas. Job Cuts Fall in September; Hiring Plans Up 3% Over 2025 on Weak Early Seasonal Hiring. October 1, 2026. Reports 120,136 announced job cuts attributed by employers to AI through September 2026. https://www.challengergray.com/blog/job-cuts-fall-in-september-hiring-plans-up-3-over-2025-on-weak-early-seasonal-hiring/ Challenger Gray Christmas
[10] Steve Jobs. Make Something Wonderful: Steve Jobs in His Own Words. Edited by Leslie Berlin, Steve Jobs Archive, 2023. See “Speech at the International Design Conference in Aspen” (1983) and “Interview with The New Yorker” (1977). https://book.stevejobsarchive.com/ Steve Jobs Archive
[11] Associated Press. Canadian Author Anne Carson Wins Literature Nobel for “Bold and Inventive” Works That Blend Genres. October 8, 2026. https://apnews.com/article/05a268ffc76827d273f9d986c5a5a65e AP News
[12] Sappho. If Not, Winter: Fragments of Sappho. Translated by Anne Carson. Alfred A. Knopf, 2002; Vintage paperback, 2003. See also Dimitrios Yatromanolakis, “Fragments, Brackets, and Poetics: On Anne Carson’s If Not, Winter,” International Journal of the Classical Tradition 11, 266–272 (2004). https://www.penguinrandomhouse.com/books/160692/if-not-winter-by-sappho-translated-by-anne-carson/ https://doi.org/10.1007/BF02720036 Penguin Random House
[13] The Secret Life of Walter Mitty. Directed by Ben Stiller, 2013. Sean O’Connell’s dialogue in the snow leopard scene; quotation listed on IMDb. https://www.imdb.com/title/tt0359950/quotes/ IMDb