EconFaithAI · An Illustrated Guide
Where technology has taken us, where it's going, and what our place in it may be.
Technology has produced abundance. The abundance is real, and worth celebrating. But the costs we are paying for it — in formation, attention, intelligence, concentration, and moral practice — are real too. This is the whole picture, in eighteen short sections.
Across nearly every measurable dimension of human flourishing, the world has improved over the last two centuries — and especially over the last hundred years. Life expectancy, GDP per capita, freedom from extreme poverty, and global literacy have all moved sharply upward in the modern era.
Real per-capita output in the United States has grown more than tenfold since 1900. Almost any good can be summoned with the click of a button. The world's information sits in our pockets. We live longer, know more, and connect more than any people who came before us.
The abundance is real, and worth celebrating. The question this guide takes seriously is what the abundance has cost — and what comes next if we do not pay attention.
Human Flourishing Index, 4000 BC – 2000 AD
Four series indexed so that 2000 AD = 100. Life expectancy, GDP per capita, % not living in extreme poverty, and global literacy — all rising sharply in the modern era.
When we consider technology's impact, we have to look beyond traditional metrics to softer, moral ones too. For most of human history, the average new technology was net-fruitful — it produced more good than it cost.
Today, the average new technology may actually be net negative on fruitfulness — things like love, joy, peace, patience, kindness, goodness, gentleness, faithfulness, and self-control. There are still ways to use technology fruitfully. But we have to shepherd it.
Abundance and fruitfulness are not the same thing.
Net Fruitfulness of New Technology, 3000 BC – 2024 AD
Rolling average across major technologies scored on the Fruit of the Spirit index. Above zero = net-positive contribution.
In the past, the most powerful forces shaping a child's life were people — parents, teachers, peers, pastors. Today, for the first time in human history, that force is an algorithm. What used to be local formation is now global. What used to be incentive-aligned may not have the child's best interest in mind.
Beginning around 2012, algorithms began shaping the formation of children — especially through their friends — to a degree that now exceeds 40% of total influence. Commensurately, we see sharp rises in negative mental health outcomes. It is important we understand these impacts.
For Gen Z, algorithmic feeds now shape a child more than school and church combined.
Share of Formative Influence on U.S. Children, 1900–2020
Seven formation channels, summed to 100%. Algorithmic-feed expansion shown in red.
U.S. Teen Depression and Anxiety Prevalence, 1900–2024
Stable for a century. Sharp inflection from 2012 onward, coinciding with smartphone saturation.
As of early 2026, over 50% of the S&P 500 market cap sits in just 17 technology-related stocks. The highest concentration ever recorded.
Unlike prior eras, this concentration is supported by real earnings and structural revenue rotation — not just elevated multiples.
A handful of firms now hold the infrastructure of intelligence itself.
Tech Share of U.S. Equity Market Value, 1900–2025
Information Technology sector plus Amazon, Alphabet, Meta, and Tesla.
A human's most valuable asset may be their attention. The average American now spends over three hours per day inside a feed-based product — software designed by some of the most sophisticated engineers in the world to keep us engaged.
Our attention is no longer ours.
Three hours a day, captured by the algorithm.
U.S. Adult Leisure Screen Time, 1900–2024
Feed-based (algorithmically curated) and non-feed-based screen time, hours per day.
Beginning around 2013, the largest tech firms shifted structurally — away from reinvesting profits into R&D and toward returning capital to shareholders via buybacks. Since then, the largest tech companies have returned more than $4 trillion to shareholders via buybacks and dividends. More than the entire GDP of Germany.
Today, the same firms are redeploying their earnings again — this time into AI infrastructure at a scale that dwarfs the buyback era. The attention-to-intelligence loop is now powering AI growth.
The attention economy funded the intelligence economy. The same firms own both.
Top-Tech Net Income and Buyback Ratio, 2000–2024
Top seven software-tech firms. Buybacks were negligible pre-2013; reached ~94% of net income by 2022.
Hyperscaler AI Capital Expenditure, 2020–2026
Combined annual CapEx by company: Microsoft, Alphabet, Amazon, Meta. 2026 figure is company guidance.
It took roughly 6,500 years for machine intelligence to grow from 0 to an IQ-equivalent of 55. It took less than four years to grow from 55 to 135. Frontier AI now beats 99% of humans on standardized reasoning benchmarks.
And it is not just one model. The entire computing-device population has migrated rightward in IQ-equivalent terms at a pace with no historical precedent. The two populations no longer share the same center.
In four years, from below functional to above genius — across an entire population.
Frontier AI IQ-Equivalent, 1900–2026
Trajectory of the leading commercial / research AI system per year. Reference lines: human average (100), genius threshold (130).
Human and Machine IQ Distributions, Animated 1990–2035
Watch the machine population (blue) migrate rightward over time as frontier capability grows.
▶ Auto-playing — 1990
Every prior major technology grew on a logistic curve — slow start, rapid growth, then natural saturation. Saturation gave our institutions time to catch up.
Today's frontier technology appears to be growing on an exponential curve — with no natural ceiling. The two curves go to completely different places.
Every prior wave saturated. This one may not.
Logistic Growth
Slow start, rapid growth, natural saturation.
Exponential Growth
Slow start, then unbounded acceleration.
The most uncomfortable property of an exponential curve is this: in the early phase, it looks indistinguishable from a logistic one. The same data points fit both models. Two observers can look at the same chart and tell two different stories — both defensible.
By the time you can see for sure which curve you are on, the trajectory is already set. We might not see it coming.
Wait-and-see is the same as decide-too-late.
Logistic vs. Exponential — Identical Until Halfway, Then Divergent
Both curves follow the same path through the first half. After x = 0.5, exponential lifts off; logistic levels off.
Past technology waves rose above institutional capacity, then plateaued. Institutions caught up. The gap closed.
Every past wave had two measurable quantities: a capability gap (how far the technology ran ahead of institutional capacity at its peak) and a catch-up time (how long it took institutions to close that gap). Both were bounded. The technology eventually plateaued; institutions eventually arrived. People got there.
An exponential wave does not plateau. The capability gap keeps growing. The catch-up time becomes undefined — there is no fixed ceiling for institutions to reach. The gap between what technology can do and what our institutions can absorb keeps widening. We may not be prepared.
Logistic waves gave institutions time. Exponential waves do not.
Institutional Capacity vs. Technology Waves
Past logistic waves (blue) rose above institutional capacity (grey) and were eventually matched. The exponential AI wave (orange) does not plateau.
Across the nine major capability dimensions analyzed in our research, humans dominate in four: reasoning under ambiguity, social and emotional embodiment, adaptation and exploration, and creativity and meaning-making.
These are not minor capabilities. They are the core work of being human — and the work most worth protecting.
The work that remains structurally human is the work most worth protecting.
Capability Dimensions Where Humans Dominate
Four of nine major capability dimensions where humans hold a clear structural advantage.
| Dimension | Advantage |
|---|---|
| Reasoning under ambiguity | Humans strong |
| Social, emotional, embodiment | Humans strong |
| Adaptation, learning, exploration | Humans strong |
| Creativity, meaning-making | Humans strong |
Machines also have clear structural advantages — particularly in throughput, memory, scale, and concurrency. The differences are not marginal. They are orders of magnitude.
The capability gap is widest where intelligence reduces to computation, recall, and parallel scale — and narrowest where intelligence requires judgment, embodiment, and meaning. The honest comparison takes both halves seriously.
The gap, where it exists, is structural — not marginal.
Specific Machine Advantages, Detail View
Snippets from the full nine-dimension comparison, focused on the areas where the machine advantage is most stark.
| Dimension | Humans | Machines (2026) |
|---|---|---|
| Working memory | 7 items | 2,000,000 items |
| Long-term storage | 2,000,000 GB | Effectively unlimited |
| Operations per second | 10 | 2,000,000,000,000+ |
| Output speed | 150 wpm | 150,000 wpm |
| Concurrent sessions | 1 | Millions |
| Geographic presence | 1 location | Everywhere at once |
| Training time | ~20 years | ~3 months |
Do the people building the technology that shapes American life represent the religious and moral makeup of the country?
The data says no. Among the top tech CEOs with publicly disclosed religious affiliation, the moral and religious composition diverges sharply from the population they serve.
The people designing the world's digital infrastructure are not drawn from the population it serves.
Religious Profile: Top Tech CEOs vs. U.S. Adult Population
Disclosed affiliations of CEOs at the largest U.S. tech companies, compared to the Pew Religious Landscape Study 2024.
Part II
The diagnosis is sobering. The response is not despair. It is engagement — deeper, closer, and more morally serious than the last generation managed.
Traditionally religious people are underrepresented in the sector that has the most impact.
The leadership gap shown in Section 13 holds at the workforce level too. The U.S. adult average for Bible-believing weekly practice is roughly 12%. Information Technology sits at 3%, Communication Services at 4%, and the AI-research sub-population at 1%.
The tech sector is roughly four times less Bible-believing than the U.S. adult average. The AI-research sub-population is twelve times less. Energy, by contrast, sits at 22%.
This is not an indictment of the people in the sector. It is a description of the gap — and a quiet call. If the people building the products that shape American life come from a population very different from the people being shaped, then morally serious builders, parents, pastors, investors, and ethicists have to engage the industry directly, not from the sidelines.
The most powerful shaping force in modern life is being built by a workforce that doesn't share the moral makeup of the people being shaped.
Bible-Believing Weekly Rate by U.S. Sector
Approximated from BLS workforce demographics × Pew religious-practice rates. Reference line at 12% = U.S. adult average. ±3 pp.
This is the most important moment of our time.
Beyond local church giving, American Christians direct an estimated $40 billion per year into mission causes — global poverty alleviation, evangelism and missions, Christian education, media, and justice work, including anti-trafficking.
But one category is almost entirely absent. Christian capital flowing into technology and AI — the very layer now shaping how children form, how adults attend, and how the country deliberates — is a rounding error against the rest.
It is hard to argue that the most influential force in modern life has earned only a rounding error of our intentional capital. If we agree that technology is shaping people, then funding the people who build it differently is a first-order moral question, not an afterthought.
If technology shapes people, technology deserves more than a rounding error of our intentional capital.
Figures exclude local-church operating giving and are directional estimates compiled from Giving USA, Empty Tomb, ECFA, and Christianity Today reporting. Treat as orders of magnitude.
Estimated Annual U.S. Christian Mission Capital, Excluding Local Church
By category, in billions of dollars per year. Directional estimates — orders of magnitude.
Financial, moral, and reach.
Capital deployed into morally serious technology is unusual among modern investments: it compounds along three different axes at the same time. Financial return, moral influence, and reach are not in tension here — they are downstream of the same well-built product. For builders, investors, and donors trying to decide where to direct attention, these are the three reasons to take this layer of the economy seriously.
Financial Impact
Annual application-layer revenue pools, today
Real revenue, real margins, real durability.
Search advertising, e-commerce, enterprise SaaS, and content platforms already account for more than a trillion dollars of annual revenue. Application-layer operators capture this value at structurally high margins — and the AI overlay is expanding, not contracting, those pools.
Moral Impact
Of children's formation now shaped by algorithmic feeds
The layer that shapes how a generation forms.
Application-layer choices — what to optimize for, what to default to, who to serve — now shape formation, attention, civic life, and intelligence itself. Building well at this layer is one of the highest-moral-leverage activities of the current moment.
Reach Impact
Reachable users per single application
A single product can touch a generation.
The marginal cost of serving one more user at the application layer is nearly zero. One well-designed product reaches billions, in hundreds of contexts, in dozens of languages, almost overnight. That kind of reach exists almost nowhere else in the modern economy.
One of the most consequential structural facts about the current moment is also one of the least understood: nearly all monetization in AI happens at the application layer, not the foundational layer. Transformers, large language models, diffusion models — the foundational breakthroughs are published, open-source, or available via API at commodity prices.
The foundational layer produces capability. The application layer — where an operator takes that capability and deploys it into a specific user context under a specific business model — is where capability becomes revenue, and where the moral architecture of the product gets built. What it is tuned for. What the defaults are. Who it serves and on what terms. The foundational technology is the same in both the engagement-extraction model and the genuine-helpfulness model. The difference is the application-layer choice the operator makes.
This is the hopeful part of the picture. The foundational capability is open, and the application layer is wide open. The cost of building has never been lower; the leverage of a single well-designed product has never been higher. For a morally serious builder, there has never been a better time to build.
Application Layer
Where revenue and moral architecture both live.
Operator chooses business model, defaults, tuning objective, and who to serve.
Foundational Layer
Published, open-source, or commodity-priced via API.
Produces capability. Does not, on its own, decide what the capability is used for.
The foundational tech is the same. The application-layer choice is everything — and the door is open.
This moment will be shaped by who shows up. The Solution Map at EconFaithAI organizes that work by role — parents, the church, builders, investors, and ethicists — with concrete actions, frameworks, and resources for each. None of this is destiny; all of it is downstream of who engages, and how.
For Parents
Shape the formation, not just the screen time.
Practical frameworks for navigating algorithmic feeds, school technology, and adolescent identity formation in an attention-saturated world.
Open the parent's roleFor the Church
Re-occupy the interior-shaping role.
Where pastors and faith communities can act as a moral thermostat — not a thermometer — in an attention economy that has displaced traditional formation channels.
Open the church's roleFor Builders
Build at the application layer with intention.
The greed-to-generosity model: what a morally serious application-layer operator actually does differently, in the actual mechanics of product, pricing, and tuning.
Open the builder's roleFor Investors
Deploy capital where moral leverage is highest.
Why the moral leverage hierarchy points toward application-layer companies — and a framework for the capital that funds the decisions that matter.
Open the investor's roleFor Policymakers
Bridge capability and care with principled policy.
Targeted, well-informed policy at the application layer — age-appropriate design, algorithmic transparency, and accountability where harm is deployed.
Open the policy levers