Pinned Tweet
Does chatGPT Dream of Electric Sheep?
“Empathy, evidently, existed only within the human community, whereas intelligence to some degree could be found throughout every phylum and order"
“Science Fiction, is an art form that paints a picture of the future.” - Anthony Merrydew
n has always enamored technologists and futurists alike. Through captive storytelling, creative world building, imaginative future technologies, and prescient writings of the future, science fiction novels have inspired us all. Certain classics have had even more impact, so accurately pinpointing the future that it would seem almost prophetic. Whether the future is just deterministic, or history always repeats itself, these authors opined on subject matters, and technologies, that we are living today.
Noah Smith recently wrote on this very subject, defending the importance, and relevance, of science fiction classics. Centering in on how these classics have inspired modern technologies, and societies, and are not just prescient warnings for future societies, but imaginative prophesies of future technologies.
One of these classics is Phillip K. Dick’s “Do Androids Dream of Electric Sheep” which was later reimagined in film as the cult classic Blade Runner. It would be pertinent to first set the stage here by giving a brief synopsis of this masterpiece. The book takes place in 1992 after World War Terminus decimated what was then the near future Earth. A post nuclear global war has destroyed the atmosphere and environment, and leads the United Nations to encourage more mass emigrations to off-world colonies to preserve humanity's genetic integrity. In order to encourage more emigration, humans are given the incentive of free personal androids if they move to a colony. The Rosen Association manufactures these androids on a colony on Mars, but some androids rebel and escape to Earth, where they hope to remain undetected given their human like appearance. American and Soviet police departments remain vigilant, and enlist android bounty-hunting officers to hunt these Androids down. On post apocalyptic Earth, owning real animals has become a fashionable status symbol, both because mass extinctions have made authentic animals rare and because of the accompanying cultural push for greater empathy. The less fortunate, however, can only afford realistic-looking robot imitations of live animals. For example, Rick Deckard, the novel's protagonist, owns a robotic black-faced sheep. The popular trend for increased empathy coincidentally motivated a new technology-based religion called Mercerism, which uses "Empathy Boxes" to link users simultaneously to a virtual reality of collective suffering, centered on a martyr-like character, Wilbur Mercer, who eternally climbs up a hill while being hit with crashing stones. Mercerism’s ultimate goal is to make humans more “human” through building a collective conscience of empathy and compassion while doing good in the community. Acquiring high-status animal pets and linking in to empathy boxes appear to be the only two ways followers of Mercerism strive for existential fulfillment.
The parallels between dystopian San Francisco, and modern day San Francisco (as well as broader 21st century American society) are astounding. It seems society today has maybe avoided nuclear war, and we haven’t yet nailed fully functioning autonomous humanoid robots, but we have gone down two other foretold dystopian paths. We are on the path of worshipping Mercerism - while losing it altogether - as well as pursuing a false technology of Empathy Boxes - while rejecting a useful one of Androids. One worldview is guilty of driving the push down these dangerous two paths, and I argue that just as Mercerism created a dystopian future, so too will Effective Altruism, Doomerism, and Wokeism do the same. Ultimately, the choice to worship a false empathy will ultimately make AGI an Empathy Box, which will lead us down a false humanity, and toward a dystopian future.
“An android…doesn’t care what happens to another android. That’s one of the indications we look for." - Do Androids Dream of Electric Sheep
Mercy
In ways, Mercerism was a prelude to where the internet has gone today. We use social media as an Empathy Box to shame society into feeling the same guilt as not just one Wilbur Mercer, but all oppressed peoples throughout history.
In fact, Woke-ism of today is eerily similar of the Mercerism of the alternate future. At one point in the novel it is stated that someone’s “schedule for today lists a six-hour self-accusatory depression”. Implying that is is righteous to accuse not just others, but one’s own self. Mercerists even believe that “empathy…must be limited to herbivores or anyhow omnivores who could depart from a meat diet. Because, ultimately, the empathic gift blurred the boundaries between hunter and victim, between the successful and the defeated.” This reads like something I’d read on X from a Gen Z activist who has decided that being vegan is the best way to fight oppression. We need to look no further than the recent version of Google Gemini that spits out images of a black George Washington, which more so seems to be the dreams of an ultra-left former citizen of CHAZ than a hallucination of a finely tuned foundation model.
With all that said, just as the humans in the novel continually pursued mercy, compassion, and empathy through technology - they continually lacked it. Our society is becoming no different. Our society is becoming a place where one is guilty and punished upon accusation. Where being human and making mistakes is not tolerable. Where a race, or an idea, is tantamount to unforgivable and original sin. Where more and more of the human experience is influenced by an authoritarian rule in both the virtual and natural realities. Where polarization has led to a loss of empathy. Where disunity is either stalling, or all out replacing unity. Ultimately, we live in a society where second chances and true mercy is seen as a form of oppression, not liberation. We must reject this notion, and not let such leanings creep into our views on technology. We must also reject this notion to not let these views creep into the very technology itself, which could be even more dangerous than where we are today.
In the grandest irony of all, the greatest benefit of an everyday, utilitarian AI will not be increased productivity or an economics of abundance or a new way of doing science—although all those will happen. The greatest benefit of the arrival of artificial intelligence is that AI’s will help define humanity. We need AI’s to tell us who we are.” - Kevin Kelly
AI: Empathy Box or Android?
No technology is being discussed more today than AGI. And it would seem the same question is being asked again. Is AGI (Artificial General Intelligence) our next Empathy Box, or is it our Android?
Within the book, we see two forms of technology, each with respective purposes. I am over generalizing here, but Androids are technology used to augment our lives as humans, and Empathy Boxes are technology used to make us find purpose in our lives as humans. There is a lesson here for us.
The Android framing is one where technology is a tool to enhance and augment human productivity and ingenuity; rapidly accelerating advancements across education, science, and more. This leads to a better society for all, continuing the arc of all technology. A more utilitarian mindset where technology, and in this case AGI, vastly augment our lives as humans. Improving humanity, not defining it. Marc Andreessen famously has his own version of Martin Luther’s 95 thesis on this very topic.
The Empathy Box framing is one where AI helps us find purpose in our lives as humans. Plugging us into a collective conscience and understanding in order to improve humanity. In the novels, the empathy boxes were simply virtual reality tech that plugged humans into a collective experience, and thus conscience. It was a technology that could be controlled and that signaled progress, that was deemed good by the masses. Androids, on the other hand, were technology that was to be feared because it could not be controlled. It was deemed bad, and hunted, not pursued. What is interesting is that for those people who view AI as the ultimate threat, they also view it as the key to finally understanding what it means to be human. A dichotomy in ways of both savior, and anti-christ. This dynamic was no more prevalent than the recent OpenAI dilemma, or the current AI Regulation Forums in DC.
“The old man said, ‘You will be required to do wrong no matter where you go. It is the basic condition of life, to be required to violate your own identity. At some time, every creature which lives must do so. It is the ultimate shadow, the defeat of creation; this is the curse at work, the curse that feeds on all life. Everywhere in the universe.” - Do Androids Dream of Electric Sheep
Conclusion
So where does this end?
For one, we have yet to see what AI will ultimately become because inherent in all these discussions are assumptions around capabilities. Most of the debates stem from either assuming unlimited capabilities, or limited, and assuming we have a role in deciding such. Either way, I think we should never view technology as something to be feared. Like all things, however, I believe technology has trended towards bettering society. AI will be no different. But I leave the debate open as to whether it has made, or will make us, more human. This leads me to my next point.
It would also seem that in the pursuit of more humanity - empathy and mercy - we are slowly losing it.
Theologians have defined biblical mercy as "not getting what you deserve”, and Scripture speaks of how “God’s kindness leads to repentance”. Our modern day Mercerists have the right idea, but the wrong application. True human empathy, seeks justice for the oppressed, and forgiveness for the oppressor. A one sided approach leads to a new religion altogether, with no transformation of the heart, and no transformation of society.
All of us would do well to remember that the same futuristic novels that predicted many of the technologies we now live in, also predicted various downfalls of societal structures, institutions, and morality. We would do well to heed the warnings of our past, and create a brighter, not bleaker, tomorrow. Let us remember what true Mercy is to attain more humanity, and let us not rely on technology like AGI as an Empathy Box, but build technology as an Android to build a better society.
“So speak and so do as those who will be judged by the law of liberty. For judgment is without mercy to the one who has shown no mercy. Mercy triumphs over judgment.” - James 2:12-13
“Empathy, evidently, existed only within the human community, whereas intelligence to some degree could be found throughout every phylum and order"
“Science Fiction, is an art form that paints a picture of the future.” - Anthony Merrydew
n has always enamored technologists and futurists alike. Through captive storytelling, creative world building, imaginative future technologies, and prescient writings of the future, science fiction novels have inspired us all. Certain classics have had even more impact, so accurately pinpointing the future that it would seem almost prophetic. Whether the future is just deterministic, or history always repeats itself, these authors opined on subject matters, and technologies, that we are living today.
Noah Smith recently wrote on this very subject, defending the importance, and relevance, of science fiction classics. Centering in on how these classics have inspired modern technologies, and societies, and are not just prescient warnings for future societies, but imaginative prophesies of future technologies.
One of these classics is Phillip K. Dick’s “Do Androids Dream of Electric Sheep” which was later reimagined in film as the cult classic Blade Runner. It would be pertinent to first set the stage here by giving a brief synopsis of this masterpiece. The book takes place in 1992 after World War Terminus decimated what was then the near future Earth. A post nuclear global war has destroyed the atmosphere and environment, and leads the United Nations to encourage more mass emigrations to off-world colonies to preserve humanity's genetic integrity. In order to encourage more emigration, humans are given the incentive of free personal androids if they move to a colony. The Rosen Association manufactures these androids on a colony on Mars, but some androids rebel and escape to Earth, where they hope to remain undetected given their human like appearance. American and Soviet police departments remain vigilant, and enlist android bounty-hunting officers to hunt these Androids down. On post apocalyptic Earth, owning real animals has become a fashionable status symbol, both because mass extinctions have made authentic animals rare and because of the accompanying cultural push for greater empathy. The less fortunate, however, can only afford realistic-looking robot imitations of live animals. For example, Rick Deckard, the novel's protagonist, owns a robotic black-faced sheep. The popular trend for increased empathy coincidentally motivated a new technology-based religion called Mercerism, which uses "Empathy Boxes" to link users simultaneously to a virtual reality of collective suffering, centered on a martyr-like character, Wilbur Mercer, who eternally climbs up a hill while being hit with crashing stones. Mercerism’s ultimate goal is to make humans more “human” through building a collective conscience of empathy and compassion while doing good in the community. Acquiring high-status animal pets and linking in to empathy boxes appear to be the only two ways followers of Mercerism strive for existential fulfillment.
The parallels between dystopian San Francisco, and modern day San Francisco (as well as broader 21st century American society) are astounding. It seems society today has maybe avoided nuclear war, and we haven’t yet nailed fully functioning autonomous humanoid robots, but we have gone down two other foretold dystopian paths. We are on the path of worshipping Mercerism - while losing it altogether - as well as pursuing a false technology of Empathy Boxes - while rejecting a useful one of Androids. One worldview is guilty of driving the push down these dangerous two paths, and I argue that just as Mercerism created a dystopian future, so too will Effective Altruism, Doomerism, and Wokeism do the same. Ultimately, the choice to worship a false empathy will ultimately make AGI an Empathy Box, which will lead us down a false humanity, and toward a dystopian future.
“An android…doesn’t care what happens to another android. That’s one of the indications we look for." - Do Androids Dream of Electric Sheep
Mercy
In ways, Mercerism was a prelude to where the internet has gone today. We use social media as an Empathy Box to shame society into feeling the same guilt as not just one Wilbur Mercer, but all oppressed peoples throughout history.
In fact, Woke-ism of today is eerily similar of the Mercerism of the alternate future. At one point in the novel it is stated that someone’s “schedule for today lists a six-hour self-accusatory depression”. Implying that is is righteous to accuse not just others, but one’s own self. Mercerists even believe that “empathy…must be limited to herbivores or anyhow omnivores who could depart from a meat diet. Because, ultimately, the empathic gift blurred the boundaries between hunter and victim, between the successful and the defeated.” This reads like something I’d read on X from a Gen Z activist who has decided that being vegan is the best way to fight oppression. We need to look no further than the recent version of Google Gemini that spits out images of a black George Washington, which more so seems to be the dreams of an ultra-left former citizen of CHAZ than a hallucination of a finely tuned foundation model.
With all that said, just as the humans in the novel continually pursued mercy, compassion, and empathy through technology - they continually lacked it. Our society is becoming no different. Our society is becoming a place where one is guilty and punished upon accusation. Where being human and making mistakes is not tolerable. Where a race, or an idea, is tantamount to unforgivable and original sin. Where more and more of the human experience is influenced by an authoritarian rule in both the virtual and natural realities. Where polarization has led to a loss of empathy. Where disunity is either stalling, or all out replacing unity. Ultimately, we live in a society where second chances and true mercy is seen as a form of oppression, not liberation. We must reject this notion, and not let such leanings creep into our views on technology. We must also reject this notion to not let these views creep into the very technology itself, which could be even more dangerous than where we are today.
In the grandest irony of all, the greatest benefit of an everyday, utilitarian AI will not be increased productivity or an economics of abundance or a new way of doing science—although all those will happen. The greatest benefit of the arrival of artificial intelligence is that AI’s will help define humanity. We need AI’s to tell us who we are.” - Kevin Kelly
AI: Empathy Box or Android?
No technology is being discussed more today than AGI. And it would seem the same question is being asked again. Is AGI (Artificial General Intelligence) our next Empathy Box, or is it our Android?
Within the book, we see two forms of technology, each with respective purposes. I am over generalizing here, but Androids are technology used to augment our lives as humans, and Empathy Boxes are technology used to make us find purpose in our lives as humans. There is a lesson here for us.
The Android framing is one where technology is a tool to enhance and augment human productivity and ingenuity; rapidly accelerating advancements across education, science, and more. This leads to a better society for all, continuing the arc of all technology. A more utilitarian mindset where technology, and in this case AGI, vastly augment our lives as humans. Improving humanity, not defining it. Marc Andreessen famously has his own version of Martin Luther’s 95 thesis on this very topic.
The Empathy Box framing is one where AI helps us find purpose in our lives as humans. Plugging us into a collective conscience and understanding in order to improve humanity. In the novels, the empathy boxes were simply virtual reality tech that plugged humans into a collective experience, and thus conscience. It was a technology that could be controlled and that signaled progress, that was deemed good by the masses. Androids, on the other hand, were technology that was to be feared because it could not be controlled. It was deemed bad, and hunted, not pursued. What is interesting is that for those people who view AI as the ultimate threat, they also view it as the key to finally understanding what it means to be human. A dichotomy in ways of both savior, and anti-christ. This dynamic was no more prevalent than the recent OpenAI dilemma, or the current AI Regulation Forums in DC.
“The old man said, ‘You will be required to do wrong no matter where you go. It is the basic condition of life, to be required to violate your own identity. At some time, every creature which lives must do so. It is the ultimate shadow, the defeat of creation; this is the curse at work, the curse that feeds on all life. Everywhere in the universe.” - Do Androids Dream of Electric Sheep
Conclusion
So where does this end?
For one, we have yet to see what AI will ultimately become because inherent in all these discussions are assumptions around capabilities. Most of the debates stem from either assuming unlimited capabilities, or limited, and assuming we have a role in deciding such. Either way, I think we should never view technology as something to be feared. Like all things, however, I believe technology has trended towards bettering society. AI will be no different. But I leave the debate open as to whether it has made, or will make us, more human. This leads me to my next point.
It would also seem that in the pursuit of more humanity - empathy and mercy - we are slowly losing it.
Theologians have defined biblical mercy as "not getting what you deserve”, and Scripture speaks of how “God’s kindness leads to repentance”. Our modern day Mercerists have the right idea, but the wrong application. True human empathy, seeks justice for the oppressed, and forgiveness for the oppressor. A one sided approach leads to a new religion altogether, with no transformation of the heart, and no transformation of society.
All of us would do well to remember that the same futuristic novels that predicted many of the technologies we now live in, also predicted various downfalls of societal structures, institutions, and morality. We would do well to heed the warnings of our past, and create a brighter, not bleaker, tomorrow. Let us remember what true Mercy is to attain more humanity, and let us not rely on technology like AGI as an Empathy Box, but build technology as an Android to build a better society.
“So speak and so do as those who will be judged by the law of liberty. For judgment is without mercy to the one who has shown no mercy. Mercy triumphs over judgment.” - James 2:12-13
9
2
55
39,543
RT @Gwynne_Shotwell: Robert and I are thrilled that our gift of over 2 million shares of SpaceX stock will be distributed to children’s Tru…
0
1,511
0
7
Christian Garrett retweeted
Men only want one thing and it's...
a neo-defense-prime-branded magnesium alloy USA-made military-grade gameboy
a neo-defense-prime-branded magnesium alloy USA-made military-grade gameboy
265
257
6,443
141,684
Christian Garrett retweeted
Neuralink’s first product is called Telepathy.
It enables high-speed communication with computers or humans simply by thinking. Output bandwidth will vastly exceed the fastest human speech or typing in a few years. Read that sentence again.
It enables high-speed communication with computers or humans simply by thinking. Output bandwidth will vastly exceed the fastest human speech or typing in a few years. Read that sentence again.
Our engineers recently visited one of our speech restoration study participants for a 72-hour hackathon.
The goal: rapidly prototype as many features as possible to enhance the experience of using Neuralink to communicate with thought.
Here are some of the things we built 🧵
3,332
6,147
49,620
6,265,372
Christian Garrett retweeted
@Walmart has spent more than 60 years helping Americans save money and live better. We built Base to do the same for home energy.
Today, Walmart and Base announced a strategic agreement and investment in Base. Starting Monday, Base’s energy products will be available at nd Sam’s Club locations.
corporate.walmart.com/news/2026/10/W…
Today, Walmart and Base announced a strategic agreement and investment in Base. Starting Monday, Base’s energy products will be available at nd Sam’s Club locations.
corporate.walmart.com/news/2026/10/W…
7
24
300
60,480
Christian Garrett retweeted
JUST IN: Google successfully activates its first AI data center satellite in space, using an onboard AI model to generate poetry while orbiting Earth.
175
532
6,717
289,756
Christian Garrett retweeted
America’s Power Company, now at America’s largest retailer.
Few companies have done more to lower Americans’ everyday bills than Walmart. The power bill is next.
Proud to partner with the @Walmart team.
Few companies have done more to lower Americans’ everyday bills than Walmart. The power bill is next.
Proud to partner with the @Walmart team.
@Walmart has spent more than 60 years helping Americans save money and live better. We built Base to do the same for home energy.
Today, Walmart and Base announced a strategic agreement and investment in Base. Starting Monday, Base’s energy products will be available at nd Sam’s Club locations.
corporate.walmart.com/news/2026/10/W…
31
24
473
48,703
Christian Garrett retweeted
edward coristine is a young hero. he is being targeted by authoritarian misandrists like kara swisher because of this, not in spite of this. piratewires.com/p/age-of-balls
16
63
962
68,957
RT @MarceloLima: There’s no way any human read this and thought, “Yeah, it’s time to dump” within 3 minutes of the email arriving.
$AMZN…
$AMZN…
0
16
0
16
Christian Garrett retweeted
Thrive’s proposed acquisition of ADT feels like 4D chess with clear downside protection, levers to increase EBITDA and many upside cases.
Here’s how I imagine @JoshuaKushner could be thinking about it.
1. Buy millions of sticky customer relationships.
ADT ended 2025 with approximately 6.1 million monitoring subscribers. As of June 2026, it generated $360M in recurring monthly revenue, equivalent to a $4.32B annual run rate, with annual gross revenue attrition of 13.1%.
Equipment is installed in the home. Customers pay every month. Families depend on the service. There are real reasons to stay.
2. Acquire a household brand people already trust.
ADT has spent more than 150 years building its name in security. Something that Warren Buffett has commonly looked for in his investments.
That could be incredibly valuable as AI moves into our homes. A startup asking for access to your cameras, front door, or family’s daily routines has to earn that trust from scratch.
3. Use AI to materially lower the cost to serve.
ADT reported $433M in customer-service expenses in 2025. Billing, scheduling, account changes, and routine troubleshooting create a substantial opportunity for automation.
A hypothetical 50% net reduction would save approximately $216M annually. At 80%, that becomes $346M.
Against the announced $15.3B enterprise value and $2.68B of 2025 adjusted EBITDA, those scenarios take the effective purchase multiple from approximately 5.7x to 5.3x or 5.1x.
4. Make customer acquisition more efficient.
ADT reported approximately $98M in advertising expenses in 2025. Google’s case studies are finding large branding lowering cost per acquisition by 22%+ using AI Max.
ADT already uses AI in marketing. Better targeting, creative testing, and lead conversion could help it acquire more customers with the same budget.
5. Own distribution for the next generation of AI hardware.
ADT already has approximately 3,100 installation and service technicians and 130 sales and service offices.
Building great hardware is hard. Getting someone to buy it, install it, connect it, and support it is another entire business.
ADT has infrastructure for all of that.
6. Expand what each household can spend with you.
The upside could extend well beyond selling another camera.
Drone security for larger properties. Intelligent perimeter systems. Eldercare monitoring. Home robots with recurring service plans.
These could carry substantial upfront order values and generate years of subscription, maintenance, and upgrade revenue.
Further out, a household agent could coordinate repairs, book service providers, and purchase supplies with the owner’s approval. That potentially adds transaction revenue to the relationship.
The bet: AI can increase lifetime gross profit per household while reducing the cost to serve it.
7. Have a potential macro tailwind.
If civil unrest or heightened concern about personal safety increases demand for protection, it could accelerate the pull for more capable security systems.
I wouldn’t say you need that scenario for the thesis to work. But it could increase the urgency to buy and willingness to pay.
8. Give it time to compound.
Thrive’s permanent capital structure supports a long ownership horizon.
Lower service costs can fund better products. Better products can increase household spending and strengthen retention. The permanent capital structure enables them to reinvest more in growth going into a macro cycle where security might be more top of might for people as well...and the existing business provides a foundation while the other new categories develop.
In the meantime Josh is buying a trusted route into millions of homes just as the range of valuable things you can sell into those homes is about to expand dramatically.
Here’s how I imagine @JoshuaKushner could be thinking about it.
1. Buy millions of sticky customer relationships.
ADT ended 2025 with approximately 6.1 million monitoring subscribers. As of June 2026, it generated $360M in recurring monthly revenue, equivalent to a $4.32B annual run rate, with annual gross revenue attrition of 13.1%.
Equipment is installed in the home. Customers pay every month. Families depend on the service. There are real reasons to stay.
2. Acquire a household brand people already trust.
ADT has spent more than 150 years building its name in security. Something that Warren Buffett has commonly looked for in his investments.
That could be incredibly valuable as AI moves into our homes. A startup asking for access to your cameras, front door, or family’s daily routines has to earn that trust from scratch.
3. Use AI to materially lower the cost to serve.
ADT reported $433M in customer-service expenses in 2025. Billing, scheduling, account changes, and routine troubleshooting create a substantial opportunity for automation.
A hypothetical 50% net reduction would save approximately $216M annually. At 80%, that becomes $346M.
Against the announced $15.3B enterprise value and $2.68B of 2025 adjusted EBITDA, those scenarios take the effective purchase multiple from approximately 5.7x to 5.3x or 5.1x.
4. Make customer acquisition more efficient.
ADT reported approximately $98M in advertising expenses in 2025. Google’s case studies are finding large branding lowering cost per acquisition by 22%+ using AI Max.
ADT already uses AI in marketing. Better targeting, creative testing, and lead conversion could help it acquire more customers with the same budget.
5. Own distribution for the next generation of AI hardware.
ADT already has approximately 3,100 installation and service technicians and 130 sales and service offices.
Building great hardware is hard. Getting someone to buy it, install it, connect it, and support it is another entire business.
ADT has infrastructure for all of that.
6. Expand what each household can spend with you.
The upside could extend well beyond selling another camera.
Drone security for larger properties. Intelligent perimeter systems. Eldercare monitoring. Home robots with recurring service plans.
These could carry substantial upfront order values and generate years of subscription, maintenance, and upgrade revenue.
Further out, a household agent could coordinate repairs, book service providers, and purchase supplies with the owner’s approval. That potentially adds transaction revenue to the relationship.
The bet: AI can increase lifetime gross profit per household while reducing the cost to serve it.
7. Have a potential macro tailwind.
If civil unrest or heightened concern about personal safety increases demand for protection, it could accelerate the pull for more capable security systems.
I wouldn’t say you need that scenario for the thesis to work. But it could increase the urgency to buy and willingness to pay.
8. Give it time to compound.
Thrive’s permanent capital structure supports a long ownership horizon.
Lower service costs can fund better products. Better products can increase household spending and strengthen retention. The permanent capital structure enables them to reinvest more in growth going into a macro cycle where security might be more top of might for people as well...and the existing business provides a foundation while the other new categories develop.
In the meantime Josh is buying a trusted route into millions of homes just as the range of valuable things you can sell into those homes is about to expand dramatically.
6
8
141
24,780
Christian Garrett retweeted
1/
Two years ago, I published The Defense Reformation. It was a warning: America was in an undeclared state of emergency and our defense industrial base wasn’t ready to meet the moment. The response I received from people in industry and government was overwhelming. Everyone who was paying attention knew the situation was bad—and they were fired up to fix it.
We’re still in a state of emergency, but there’s good news: the hard work of reform and mobilization has begun.
The founders and funders have shown up. The heretical heroes have momentum at the Pentagon. The Last Supper of defense consolidation is over and the First Breakfast has begun. New entrants are flooding the space and forcing legacy contractors to up their game.
America is entering a virtuous cycle of competition, iteration, and innovation. The Arsenal of Democracy is coming to life.
To celebrate two years of hard-fought progress, I’m giving away 180 free copies of my book with Madeline Hart, Mobilize: How to Reboot the American Industrial Base and Stop World War III.
Claim yours, read it, spread the good word, and get back in the fight:
18theses.com/#giveaway
Two years ago, I published The Defense Reformation. It was a warning: America was in an undeclared state of emergency and our defense industrial base wasn’t ready to meet the moment. The response I received from people in industry and government was overwhelming. Everyone who was paying attention knew the situation was bad—and they were fired up to fix it.
We’re still in a state of emergency, but there’s good news: the hard work of reform and mobilization has begun.
The founders and funders have shown up. The heretical heroes have momentum at the Pentagon. The Last Supper of defense consolidation is over and the First Breakfast has begun. New entrants are flooding the space and forcing legacy contractors to up their game.
America is entering a virtuous cycle of competition, iteration, and innovation. The Arsenal of Democracy is coming to life.
To celebrate two years of hard-fought progress, I’m giving away 180 free copies of my book with Madeline Hart, Mobilize: How to Reboot the American Industrial Base and Stop World War III.
Claim yours, read it, spread the good word, and get back in the fight:
18theses.com/#giveaway
31
54
403
205,669
Christian Garrett retweeted
Early Neuralink speech restoration data:
We are enabling speech to text and voice at conversational rates in participants with ALS. One participant achieved 75 WPM at 5.8% WER on a 125k-word vocabulary in their first week.
We are enabling speech to text and voice at conversational rates in participants with ALS. One participant achieved 75 WPM at 5.8% WER on a 125k-word vocabulary in their first week.
307
1,088
10,408
1,159,595
Christian Garrett retweeted
My key takeaways from the $GOOGL (Alphabet) Q3 2026 earnings call.
1. Search slowed while query growth hit a five-year high. Search and other ads grew 12% to $63.3B, down from Q2. The deceleration has four parts. FX was the biggest drag. The EU ended its de minimis exemption (duty-free treatment for cheap parcels), which hit APAC retail advertisers. World Cup media budgets landed in June instead of July. And AI Mode queries grew fast on a surface with "lower monetization" than classic Search. The first three wash out over time. The fourth is structural: more usage on a surface that earns less per query.
2. The EPS miss is mostly a SpaceX mark. Revenue grew 25% to $127.6B. Operating income grew 30% to $40.6B, a 31.8% margin. EPS came in at $2.12 against $3.02 expected because OI&E (other income and expense, the non-operating line where investment gains and losses land) swung to negative $8.2B. Most of that was an unrealized loss on the SpaceX stake after its share price fell. My rough math: after tax, that mark costs around $0.55 a share, so it explains more than half the $0.90 gap but not all of it.
3. Q4 Search is already guided softer. Lapping last year's Search acceleration cost about 3 points of growth in Q3 and will cost a couple more in Q4. FX gets slightly worse at current spot rates, mostly in Search and YouTube. The rest of Services won't make up the difference: YouTube ads grew only 9% to $11.2B, and Network grew 1% to $7.5B.
4. The AI Mode ad fix is rolling out but not proven. Ads in AI Mode go international for the first time, and multi-turn ads (ads that stay in a back-and-forth conversation) launch in the U.S. The evidence offered: inside AI Mode, sponsored and organic product listings get similar click-through rates. AI Mode queries run about 3x longer than classic searches, and most are incremental, meaning they would not have happened otherwise. What's missing is any number on how far AI Mode revenue per query trails classic Search. Agentic commerce (AI agents that shop and buy for users) is called additive, and consumer agents arrive in the Gemini app later this year. No revenue is attached to it yet.
5. Cloud doubled. Revenue grew 103% to $30.7B. Operating income was $11.1B, more than triple last year. Margin was 36.2% versus 23.7% a year ago. Backlog (signed contracts not yet booked as revenue) reached $561B, and just over 55% of it, about $310B, converts within 24 months. Two caveats. TPU system sales contributed to the growth but were not sized, and Cloud growth excluding TPUs was described as very strong with no number attached. Also, the rented third-party compute capacity came online only at the very end of Q3, so its margin drag lands in Q4.
6. TPU economics stay private. A TPU is Google's in-house AI chip, sold here as full racks with networking, storage and software. One analyst framed a TPU system at roughly $35B per gigawatt and asked how much of that becomes revenue and at what margin. Neither got an answer, because product-level financials are not disclosed. The cash cycle is clear. A signed contract shows up in backlog, Google builds inventory, and cash arrives on delivery. Inventory is only built against committed contracts. There are nearly 100 TPU customers, including Anthropic (directly and through Broadcom) and Apple, which is scaling pretraining on TPUs. New distribution comes through Oracle and Crux AI, a joint venture with Blackstone. The meaningful revenue ramp is 2027.
7. Free cash flow went negative. Q3 FCF was negative $9.1B on $51.4B of capex. Operating cash flow of $42.3B also absorbed the TPU inventory build and a $5.2B payment on a 2018 EC fine. Trailing twelve-month FCF is $19.7B. The 2026 capex guide is unchanged at $195B to $205B, and 2027 rises significantly, with numbers on the Q4 call. On top of that sit $924B in future purchase commitments and $152B in leases that haven't started and other obligations. Those are contractual maximums, and some have volume flexibility.
8. The return math is stated, not shown. Return on invested capital is said to stay meaningfully above the cost of capital, with no figure. Compute allocation is described as a long-range plan run on ROIC discipline, with no ranking of who gets capacity first. Cloud says it could serve a lot more demand with more capacity. About 65% of Q3 infrastructure spend went to servers and networking depreciated over six years, while TPU 8t claims up to 2.7x the training performance of the prior generation. My view: six years is a long life for hardware that loses that much ground each generation. Shorter lives would mean higher depreciation on top of the capex.
9. First-party model APIs process nearly 32B tokens (units of text an AI model reads or writes) per minute, up 40% QoQ and more than 4x YoY. In its first week on developer platforms, Gemini 4 Argon ramped 3.5x faster than Gemini 3, even with capacity limits. Customers who used Gemini Enterprise in Q3 2025 have more than quadrupled their spend. The Gemini app is at 1B users. This is the clearest evidence that the capex has real buyers.
Bottom line: The operating business beat. Revenue grew 25%, operating income grew 30%, Cloud doubled at a 36% margin, and the headline EPS miss is mostly a paper loss on SpaceX. The stock still fell about 10% after hours, from $322.22 to $288.35, because the market is now pricing cash, not growth. Quarterly FCF is negative, capex runs near $200B and rises again in 2027, commitments total about $1T, and the biggest profit pool is slowing as usage shifts to a surface that earns less per query. Management described every link between the spend and the payoff and did not put any number on them.
1. Search slowed while query growth hit a five-year high. Search and other ads grew 12% to $63.3B, down from Q2. The deceleration has four parts. FX was the biggest drag. The EU ended its de minimis exemption (duty-free treatment for cheap parcels), which hit APAC retail advertisers. World Cup media budgets landed in June instead of July. And AI Mode queries grew fast on a surface with "lower monetization" than classic Search. The first three wash out over time. The fourth is structural: more usage on a surface that earns less per query.
2. The EPS miss is mostly a SpaceX mark. Revenue grew 25% to $127.6B. Operating income grew 30% to $40.6B, a 31.8% margin. EPS came in at $2.12 against $3.02 expected because OI&E (other income and expense, the non-operating line where investment gains and losses land) swung to negative $8.2B. Most of that was an unrealized loss on the SpaceX stake after its share price fell. My rough math: after tax, that mark costs around $0.55 a share, so it explains more than half the $0.90 gap but not all of it.
3. Q4 Search is already guided softer. Lapping last year's Search acceleration cost about 3 points of growth in Q3 and will cost a couple more in Q4. FX gets slightly worse at current spot rates, mostly in Search and YouTube. The rest of Services won't make up the difference: YouTube ads grew only 9% to $11.2B, and Network grew 1% to $7.5B.
4. The AI Mode ad fix is rolling out but not proven. Ads in AI Mode go international for the first time, and multi-turn ads (ads that stay in a back-and-forth conversation) launch in the U.S. The evidence offered: inside AI Mode, sponsored and organic product listings get similar click-through rates. AI Mode queries run about 3x longer than classic searches, and most are incremental, meaning they would not have happened otherwise. What's missing is any number on how far AI Mode revenue per query trails classic Search. Agentic commerce (AI agents that shop and buy for users) is called additive, and consumer agents arrive in the Gemini app later this year. No revenue is attached to it yet.
5. Cloud doubled. Revenue grew 103% to $30.7B. Operating income was $11.1B, more than triple last year. Margin was 36.2% versus 23.7% a year ago. Backlog (signed contracts not yet booked as revenue) reached $561B, and just over 55% of it, about $310B, converts within 24 months. Two caveats. TPU system sales contributed to the growth but were not sized, and Cloud growth excluding TPUs was described as very strong with no number attached. Also, the rented third-party compute capacity came online only at the very end of Q3, so its margin drag lands in Q4.
6. TPU economics stay private. A TPU is Google's in-house AI chip, sold here as full racks with networking, storage and software. One analyst framed a TPU system at roughly $35B per gigawatt and asked how much of that becomes revenue and at what margin. Neither got an answer, because product-level financials are not disclosed. The cash cycle is clear. A signed contract shows up in backlog, Google builds inventory, and cash arrives on delivery. Inventory is only built against committed contracts. There are nearly 100 TPU customers, including Anthropic (directly and through Broadcom) and Apple, which is scaling pretraining on TPUs. New distribution comes through Oracle and Crux AI, a joint venture with Blackstone. The meaningful revenue ramp is 2027.
7. Free cash flow went negative. Q3 FCF was negative $9.1B on $51.4B of capex. Operating cash flow of $42.3B also absorbed the TPU inventory build and a $5.2B payment on a 2018 EC fine. Trailing twelve-month FCF is $19.7B. The 2026 capex guide is unchanged at $195B to $205B, and 2027 rises significantly, with numbers on the Q4 call. On top of that sit $924B in future purchase commitments and $152B in leases that haven't started and other obligations. Those are contractual maximums, and some have volume flexibility.
8. The return math is stated, not shown. Return on invested capital is said to stay meaningfully above the cost of capital, with no figure. Compute allocation is described as a long-range plan run on ROIC discipline, with no ranking of who gets capacity first. Cloud says it could serve a lot more demand with more capacity. About 65% of Q3 infrastructure spend went to servers and networking depreciated over six years, while TPU 8t claims up to 2.7x the training performance of the prior generation. My view: six years is a long life for hardware that loses that much ground each generation. Shorter lives would mean higher depreciation on top of the capex.
9. First-party model APIs process nearly 32B tokens (units of text an AI model reads or writes) per minute, up 40% QoQ and more than 4x YoY. In its first week on developer platforms, Gemini 4 Argon ramped 3.5x faster than Gemini 3, even with capacity limits. Customers who used Gemini Enterprise in Q3 2025 have more than quadrupled their spend. The Gemini app is at 1B users. This is the clearest evidence that the capex has real buyers.
Bottom line: The operating business beat. Revenue grew 25%, operating income grew 30%, Cloud doubled at a 36% margin, and the headline EPS miss is mostly a paper loss on SpaceX. The stock still fell about 10% after hours, from $322.22 to $288.35, because the market is now pricing cash, not growth. Quarterly FCF is negative, capex runs near $200B and rises again in 2027, commitments total about $1T, and the biggest profit pool is slowing as usage shifts to a surface that earns less per query. Management described every link between the spend and the payoff and did not put any number on them.
57
111
1,294
251,889
Christian Garrett retweeted
$GOOG盘后大跌历史上这个跌幅也有3次。财报3年来第一次低于预期，特别是现在比较敏感和紧张市场环境下，这必然短期股价巨大压力。
Search和cloud，双引擎看起来一个要熄火，一个虽然加速还在烧钱，那么如何看，怎么办，投资策略和估值？
月美伊冲突大反弹后从低点270一路拉升超400。然后真买单开始，需要证明了，从AI模型研发之后、加大AI投资、股权融资、现金流转负，一个个利空哪怕巴菲特350保底，平均300的成本。
今天的财报雪上加霜，最吓人是主营现金牛“搜索”业务收入增速放缓为12%同比年度增长，预计Q4下一季度还可能继续放缓到增速10%，那么Search业务是不是在AI时代要死了。再加上本季度自由现金流负90亿大低于预期。哪怕云服务第二曲线业务，同比惊人的翻倍103%，也怀疑是不是这里面不是纯服务，而是包含了TPU的销售冲量？
再加上Q4接下来还可能Cloud外包第三方算力，因为自建速度满足不了需求，虽然是好事，但是肯定降低云服务本季度Q3的36.2%的毛利。而且Q4按照年度预算还有近600-700亿的投资。
这么看股价不跌，天理难容啊。但是基本上这么解读的都是财经小报和自媒体网红水平。
最重要信息是，本季度披露AI的资本投资回报远超资金成本，AI算力和平台需求持续性明确。按照在手订单一半以上，24月兑现，明年Cloud收入含TPU超2000亿。利润率35%以上。
Search现在确实受到AI mode的模式影响，但是这是正常的“范式转移”，本季度增速下降最大原因是汇率问题，而且相对世界杯的季度平滑，下一步随着AI广告植入和各种场景拓展，变现渠道和效率，高转化的对话多轮下，Agent时代的机会。当然不能说现在还在拐点，但是说搜索和点击广告要死了，这好比Google当初从网页转向移动端广告一样，忽视了这种巨大的范式可能性。我个人认为合理在至少维持10%增速的现金流底盘。
TPU外售虽然对现金流有存货要求负面影响，但是现在这种需求下，都是先有合同再出货直接进入客户或者第三方机房，这种收入虽然不比经常性的云收入，但是资本回报效率很高。
最悲观的就是本季度现金流大幅低于预期，本质上是因为一笔52亿的历史罚款支出。和AI的投入资本效率产出没关系，把这额外支出加回，再进一步不考虑可能的32亿库存现金流支出，实际上本季度的自由现金流差不多平衡了。当然下一季度或者近期随着TPU对外销售指数性爆发，估计还会占用现金流。
至于如何投资，大跌后买入还是卖出，合理的估值区间和仓位见订阅者专享。
Search和cloud，双引擎看起来一个要熄火，一个虽然加速还在烧钱，那么如何看，怎么办，投资策略和估值？
月美伊冲突大反弹后从低点270一路拉升超400。然后真买单开始，需要证明了，从AI模型研发之后、加大AI投资、股权融资、现金流转负，一个个利空哪怕巴菲特350保底，平均300的成本。
今天的财报雪上加霜，最吓人是主营现金牛“搜索”业务收入增速放缓为12%同比年度增长，预计Q4下一季度还可能继续放缓到增速10%，那么Search业务是不是在AI时代要死了。再加上本季度自由现金流负90亿大低于预期。哪怕云服务第二曲线业务，同比惊人的翻倍103%，也怀疑是不是这里面不是纯服务，而是包含了TPU的销售冲量？
再加上Q4接下来还可能Cloud外包第三方算力，因为自建速度满足不了需求，虽然是好事，但是肯定降低云服务本季度Q3的36.2%的毛利。而且Q4按照年度预算还有近600-700亿的投资。
这么看股价不跌，天理难容啊。但是基本上这么解读的都是财经小报和自媒体网红水平。
最重要信息是，本季度披露AI的资本投资回报远超资金成本，AI算力和平台需求持续性明确。按照在手订单一半以上，24月兑现，明年Cloud收入含TPU超2000亿。利润率35%以上。
Search现在确实受到AI mode的模式影响，但是这是正常的“范式转移”，本季度增速下降最大原因是汇率问题，而且相对世界杯的季度平滑，下一步随着AI广告植入和各种场景拓展，变现渠道和效率，高转化的对话多轮下，Agent时代的机会。当然不能说现在还在拐点，但是说搜索和点击广告要死了，这好比Google当初从网页转向移动端广告一样，忽视了这种巨大的范式可能性。我个人认为合理在至少维持10%增速的现金流底盘。
TPU外售虽然对现金流有存货要求负面影响，但是现在这种需求下，都是先有合同再出货直接进入客户或者第三方机房，这种收入虽然不比经常性的云收入，但是资本回报效率很高。
最悲观的就是本季度现金流大幅低于预期，本质上是因为一笔52亿的历史罚款支出。和AI的投入资本效率产出没关系，把这额外支出加回，再进一步不考虑可能的32亿库存现金流支出，实际上本季度的自由现金流差不多平衡了。当然下一季度或者近期随着TPU对外销售指数性爆发，估计还会占用现金流。
至于如何投资，大跌后买入还是卖出，合理的估值区间和仓位见订阅者专享。
34
23
345
191,435
Christian Garrett retweeted
Lol. Google so down 10% because…
- “Search revenue decelerated”… but overall search engagement is way up. It’s because AI Mode isn’t being monetized yet.
- They invested a lot in more data centers.
- Their SpaceX stock had a temporarily loss.
- Google Cloud is at 105% YoY growth.
This is going to recover so fast…
- “Search revenue decelerated”… but overall search engagement is way up. It’s because AI Mode isn’t being monetized yet.
- They invested a lot in more data centers.
- Their SpaceX stock had a temporarily loss.
- Google Cloud is at 105% YoY growth.
This is going to recover so fast…
27
11
491
32,451
Christian Garrett retweeted
Wait wait wait. The pretzel braining “Anthropic will never be profitable” crowd are now saying that the compute Anthropic gets from SpaceX is at a DISCOUNT?!?!
I mean common! Anyone paying half attention knows that the SpaceX compute comes at a massive premium. SpaceX compute may very well be the most EXPENSIVE compute
There’s honest debate, and then there’s straight up lunacy. Guess he’s got a paid newsletter to sell….
I mean common! Anyone paying half attention knows that the SpaceX compute comes at a massive premium. SpaceX compute may very well be the most EXPENSIVE compute
There’s honest debate, and then there’s straight up lunacy. Guess he’s got a paid newsletter to sell….
“Is this just a big circle jerk?” Jon welcomes Ed Zitron (@edzitron), host of the podcast Better Offline, to make sense of the financial house of cards propping up AI and big tech. New pod out now!
41
14
326
33,604
Christian Garrett retweeted
LIFE returns to print 11.23.26.
Read our Editors’ Letter and learn more about the new LIFE here: new.life.com
Read our Editors’ Letter and learn more about the new LIFE here: new.life.com
104
1,108
4,275
378,618