An interactive reference
AI is moving fast. The economy is slower.
For 150 years, US living standards rose ~2% a year, through electricity, the transistor, and the internet. AI may be another general-purpose technology. The question is how fast task-level progress becomes measured growth, and how much depends on the weak links still handled by people.
I built this on Charles I. Jones's weak-links view of AI and growth, especially Jones & Tonetti's automation forecast, and set it against the wider recent evidence on how fast AI is actually moving: the capability curve, the 2% line, computers' shrinking GDP share, weak-link business chains, the spectrum of expert forecasts, the objections I find hardest to answer, and role horizons I update as the evidence comes in.
Source: The 150-year ~2% growth figure and forward paths recreate Jones & Tonetti (2026) and Jones's 2026 AI/growth talk; the earlier Aghion, Jones & Jones (2019) model is the baseline. Underlying income history approximated from Maddison/BEA for display. · Jones & Tonetti (2026), “Past Automation and Future A.I.” · “A.I. and Our Economic Future,” Professor Chad Jones (Stanford GSB) · Aghion, Jones & Jones (2019)
The big idea
A chain is only as strong as its weakest link.
Most businesses are chains of tasks, handoffs, approvals, and accountability. Make a few links cheap and the total result can still be capped by the remaining links. Pick a business below and change the tasks AI can already do.
To ship a product feature, this business runs 9 tasks in a chain. AI makes the routine ones strong; the human judgment, accountability, and physical tasks stay weak.
Each bar is a real task. Click to automate it to 100%.
Overall chain strength
59%
0 of 9 tasks automated
Getting stronger. Keep raising the lowest tasks, not the ones that are already high.
Want to find the weak links in your own team's work? Run a weak-link audit with the people who do it.
Source: Interactive business chains are site-authored illustrative composites. The weak-link production framework comes from Jones (2011), and the AI automation application follows Jones & Tonetti (2026). · Jones (2011), AEJ: Macroeconomics · Jones & Tonetti (2026), “Past Automation and Future A.I.”
A billion times the compute
The compute a dollar buys has doubled every couple of years for decades; by now it is roughly a billion times what the mainframe era got. Productivity growth still sits near 2%. The cheap input is not always the scarce input. Judgment, attention, and trust often still gate the result.
Free doesn't mean infinite
In Jones's infinite-automation result, driving a task's cost to zero gives a finite level gain, captured by 1/(1-s), unless the remaining weak links move too. If software is a few percent of GDP, infinitely cheap software makes us a few percent richer, once. Lasting growth means moving the next bottleneck, and the next.
The puzzle
Transformative technologies, and still 2% a year.
Electricity, internal combustion, antibiotics, semiconductors, and the internet all changed everyday life. Measured growth still stayed close to its long-run path: each wave kept the 2% line going as the previous one ran out of steam.
Source: US real income per person, 2025 dollars, approximated for display. The 150-year roughly 2% growth framing follows Jones's AI/growth work and Jones & Tonetti (2026). · Jones, “A.I. and Our Economic Future” (forthcoming, JEP) · Jones & Tonetti (2026), “Past Automation and Future A.I.” · Maddison Project
The same 2%, around the world
At the frontier, income per person has grown about 2% a year in most decades since 1870, and more slowly since 2000. Countries far behind can grow much faster for a while by adopting what already works elsewhere, then slow as they close the gap: Japan and South Korea show the full arc, while China and India are still on the way up. Catch-up growth doesn't mean the frontier itself has sped up.
Each point is the average yearly growth of real income per person over the previous 20 years. Chips show the 2002–2022 average.
Source: Real GDP per person in 2011 international dollars, accessed via Our World in Data; growth is the average over the previous 20 years. Early years are sparse for some countries. · Maddison Project Database, version 2023 (Bolt and van Zanden, 2024, Journal of Economic Surveys)
Computers are everywhere, but their price falls faster than their quantity rises. The weak links capture the value.
Source: Adapted from Jones & Tonetti (2026) and Jones's 2026 AI/growth talk; computer and electronics value-added pattern approximated from BEA/BLS data for display. · Jones & Tonetti (2026), “Past Automation and Future A.I.” · “A.I. and Our Economic Future,” Professor Chad Jones (Stanford GSB) · BEA · BLS
The chart that surprised me
Computers are everywhere, so I assumed their slice of GDP had grown. It has fallen by about a third since 2000. Prices dropped faster than we bought more, so the abundant input got cheap while the scarce inputs, people, trust, coordination, kept most of the value.
That is the weak-link pattern in one chart: making an input cheap is not the same as automating the work it sits inside.
The capability curve
The models really are getting faster.
None of this means the capability is standing still. METR measures the length of task an AI agent can finish on its own, and that horizon has doubled about every 6.2 months since 2019, and about every 4.2 months since 2023. Capability is on a genuine exponential. The weak-link claim is not that AI is slow, it is that fast capability meets slow diffusion, and the gap between the two curves is where the remaining human links live.
Source: Task-completion horizons use METR's public Time Horizon 1.1 YAML (p50 estimates for SOTA frontier points). METR notes that measurements above 16 hours are unreliable with the current task suite, so the dashed forward trend is illustrative. Compute and cost context from Epoch AI and the Stanford AI Index. · METR Time Horizon 1.1 dashboard · Epoch AI (2024), Can AI Scaling Continue Through 2030? · Stanford HAI, 2025 AI Index
Three of these clocks run in months and one runs in years. That mismatch is the whole argument: the inputs are sprinting while measured growth keeps near its 2% pace, because output still waits on the links that have not been automated.
Where we are today
Dates, predictions, and bottlenecks.
The page tracks claims that can be checked: adoption milestones, failed forecasts, and places where the remaining bottleneck is still human. The pattern so far: adoption can be very fast while economic transformation is slower.
Milestone timeline
- Mar 2004Verified
DARPA Grand Challenge: zero finishers
Not a single autonomous vehicle completed the desert course. The starting gun for self-driving, and a reminder of how hard the physical world is.
- Oct 2005Verified
Stanford's "Stanley" wins the DARPA Grand Challenge
One year after zero finishers, Sebastian Thrun's team completed the 132-mile course. Twenty years later, robotaxis are still rare outside a few cities.
- Sep 2012Verified
AlexNet ignites the deep-learning era
A deep neural network crushed the ImageNet benchmark, kicking off the modern wave of AI capability gains.
- Mar 2016Verified
AlphaGo defeats Lee Sedol
DeepMind's system beat a top human Go player 4–1, years ahead of expert expectations for the game.
- Jun 2020Verified
GPT-3 released as a developer API
A 175B-parameter model showed broad few-shot ability, but reached developers, not a mass consumer audience.
- Nov 2020Verified
AlphaFold cracks protein structure prediction
At CASP14, AlphaFold2 reached near-experimental accuracy, transforming structural biology. The work earned a 2024 Nobel Prize in Chemistry.
- Nov 2022Verified
ChatGPT launches
OpenAI released a conversational interface over GPT-3.5. Adoption was almost immediate.
- Jan 2023Verified
ChatGPT reaches ~100M monthly users in ~2 months
The fastest-growing consumer app at the time. NOTE: this was ChatGPT (GPT-3.5), not GPT-3. Adoption can be blindingly fast even when economic transformation is slow.
~100M users in ~2 monthsReuters (Feb 2023) - Mar 2023Verified
GPT-4 released
A large multimodal model with markedly stronger reasoning and coding, the workhorse behind the first wave of AI copilots.
- Aug 2024Verified
Waymo scales paid robotaxi rides
Driverless rides became a daily reality in San Francisco and Phoenix, yet remained rare nationally. Diffusion measured in decades, not years.
- Sep 2026Verified
Claude Opus 5.5 sets a new capability record (ECI 167.35)
Anthropic's model tops the Epoch Capabilities Index, which stitches many benchmarks into one score.
- Oct 2026Projection
Today: you are here (Oct 2026)
Adoption is nearly universal and capability keeps compounding; the economy-wide productivity jump is still small. Fast in a few lanes, slow across the rest, the split the weak-link view predicts.
How fast is AI?, present-day marker
Signals to watch, not yet verified
Forward-looking markers, kept out of the verified record. Treat each as a claim until a primary source confirms it.
- Nov 2025Claim
A frontier model reportedly tops an engineering take-home exam
A leading lab's multi-hour take-home hiring exam was reportedly completed by its newest model at a score higher than any human on record. Treat as a claim pending public confirmation.
Reported, pending verification - Dec 2025Claim
Coding agents start holding context across multi-hour tasks
The frontier shifts from quick answers to agents that run for hours, the first credible long-horizon autonomy in software.
Unverified signal - Feb 2026Claim
Enterprises move AI agents into real workflows
Pilots turn into production: support, coding, and back-office agents handle a slice of real volume while humans still own the judgment calls. Adoption races ahead of measured productivity.
Unverified signal
Prediction scoreboard
“They should stop training radiologists now.”
Radiologists are still trained and hired. AI now helps read scans, yet Mayo Clinic's radiology staff alone grew 55% after 2016, to more than 400. Consults, hard cases, and sign-off stayed human.
Lesson: weak links
“…next year for sure, we'll have over a million robotaxis on the road. The fleet wakes up with an over the air update; that's all it takes.”
No Tesla robotaxi service ran in 2020. Paid rides began in Austin in June 2025 with safety monitors on board, and by mid-2026 the driverless fleet was about 20 cars.
Lesson: the physical world diffuses slowly
“…these macroeconomic effects appear nontrivial but modest—no more than a 0.66% increase in total factor productivity (TFP) over 10 years.”
Not yet due. Productivity statistics can't yet separate AI's contribution from everything else.
Lesson: the slow clock
“…a superhuman coder (SC): an AI system that can do any coding tasks that the best AGI company engineer does, while being much faster and cheaper.”
Not yet due.
Lesson: horizon is the crux
The median expert expected AI to reach gold-medal level at the International Mathematical Olympiad around 2030; the median superforecaster, around 2035.
AI systems reached gold-medal level at the July 2025 Olympiad, five years before the median expert forecast and ten before the median superforecaster's. FRI finds forecasters have repeatedly underestimated progress on benchmarks.
Lesson: capability can arrive early
Find the bottleneck, by domain
For each field: what AI has automated, and what remains hard to automate. Automating 75% of the tasks doesn't finish the job. The rest set the pace. Each card carries a real-usage signal from the Anthropic Economic Index, which maps millions of actual AI conversations onto real-world tasks.
Software engineering
The first thing being automated, and still bottlenecked by judgment and accountability.
Automated
- Autocomplete & boilerplate
- Test generation
- Bug discovery in mature code
- Routine refactors
Weak link
- System design & architecture
- Ambiguous requirements
- Integrating with messy production data
- Owning the outcome when it breaks
Real usage · Computer and mathematical tasks are the largest slice of current Claude usage (~19.8%), with software developers leading individual job-title usage.
Even if AI writes most code, integrating it into every business is a long, engineer-heavy process.
Radiology
Hinton's 2016 test case. AI got better at reading scans; the job grew anyway.
Automated
- Scan triage
- Cancer-detection assist
- Measurement & flagging
Weak link
- Surgical & treatment consults
- The hardest, ambiguous scans
- Liability and sign-off
- Talking to patients and clinicians
Real usage · Healthcare-practitioner usage is tiny in the current index; radiologic technicians register at 0.00% of mapped Claude conversations.
More radiologists in 2026 than 2016, and better paid. Weak links win.
Driving
Seemingly simple, actually decades-long. The canonical slow-diffusion case.
Automated
- Highway driving
- Mapped-city autonomy (SF, Phoenix)
- Sensing & lane control
Weak link
- Long-tail edge cases
- Unmapped regions
- Adverse weather
- Nationwide scale & trust
Real usage · Transportation and material-moving roles barely register in software-based AI usage data; many physical occupations show 0.00% usage.
20+ years from "solved in 5" to still-rare. The physical world bottlenecks.
Early-childhood teaching
A 'someday' automation gated almost entirely by trust and safety, not capability.
Automated
- Content delivery (potential)
- Practice & feedback (potential)
Weak link
- Trust & safety with children
- Care and supervision
- Parental acceptance
- Liability
Real usage · Educational instruction and library tasks are a visible usage category (~7.1%), but the index shows tutoring and content work, not classroom supervision.
We could build a world-class teaching robot before we'd let it run a kindergarten unsupervised.
Source: Domain bottlenecks are site synthesis from the weak-link framework. The self-driving cars slow-diffusion example follows Jones's discussion; real-usage signals are read from the Anthropic Economic Index; item-level timeline and prediction cards carry their own factual sources. · Jones, “A.I. and Our Economic Future” (forthcoming, JEP) · Jones & Tonetti (2026), “Past Automation and Future A.I.” · Anthropic Economic Index
Field evidence: real gains, unevenly shared
Controlled studies find large average productivity gains from generative AI, concentrated where skill was scarce. The strongest workers, the hardest links, move least. That is the weak-link pattern showing up in the labor data.
Source: Customer-support field experiment from Brynjolfsson, Li & Raymond (2025); writing experiment from Noy & Zhang (2023). Figures are the studies' reported effects. · Brynjolfsson, Li & Raymond (2025), Generative AI at Work · Noy & Zhang (2023), Science
The forecast
Calibrate the model, then run it forward.
The Jones & Tonetti weak-link automation model is calibrated to history and run forward. Even the aggressive “Moore's Law everywhere” case takes about 30 years to play out, because output depends on the links that have not been automated yet.
Source: Based on Jones & Tonetti's automation-meets-weak-links forecast model; anchor points interpolated for display. · Jones & Tonetti (2026), “Past Automation and Future A.I.” · “A.I. and Our Economic Future,” Professor Chad Jones (Stanford GSB) · Aghion, Jones & Jones (2019)
Not sure which settings match your view? Take the six-question quiz and see whose forecast yours is closest to.
Where this model sits among the experts
Jones & Tonetti are one view on a wide dial. Here is the same question, how fast AI changes growth, answered from the skeptical end to the explosive end. Hover any view to read its claim.
Daron Acemoglu
The conservative anchor: predicted TFP gains over the next 10 years are below 0.53%, with even the initial task-based estimate no more than 0.66%. Real, but small.
macro estimateAcemoglu (2024)Jones & Tonetti
Weak links cap the gains. In the extreme Moore's-Law-everywhere calibration, growth exceeds 7% by 2030 and 13% by 2040, but the path still depends on whether essential human-only tasks remain.
model calibrationJones & Tonetti (2026)Trammell & Korinek
Surveying the field: fully automating production can break the old growth regularities and raise the growth rate, but the range of plausible outcomes is genuinely wide.
survey paperTrammell & Korinek (2023)Erdil & Besiroglu
The rigorous explosive-growth case: broad automation could make growth roughly an order of magnitude faster, then the paper tests that against nine named counterarguments.
review paperErdil & Besiroglu (2023)Translate the views into comparable growth paths
These lenses convert each paper's headline claim into the same annual-growth-rate frame. The chart is a comparison tool, not a claim that the papers share one model.
Source: Lens curves are site-created translations for comparison. Acemoglu is smoothed from a 10-year TFP level estimate; Jones & Tonetti uses the continuing-the-past weak-link calibration; Trammell & Korinek is shown as a range because the paper is conditional; Erdil & Besiroglu is shown as a ramp and uncertainty band around their order-of-magnitude/explosive-growth framing and Section 3 counterarguments. · Acemoglu (2024), The Simple Macroeconomics of AI · Jones & Tonetti (2026), “Past Automation and Future A.I.” · Trammell & Korinek (2023), Economic Growth under Transformative AI · Erdil & Besiroglu (2024), Explosive Growth from AI Automation
How many IT booms is AI?
The late-1990s computer and internet boom added about 0.67 percentage points a year to US productivity growth for roughly a decade. Here are AI forecasts in that unit: 1× would be another IT boom.
- The IT boom (1995–2004)
Oliner & Sichel (2000)
1.0×
- Acemoglu (2024)
at most 0.66% more total factor productivity over ten years
<0.1×
- Goldman Sachs (2023)
productivity growth 1.5 points higher for ten years after wide adoption
2.2×
- Skeptical: Daron Acemoglu
as drawn on this page
<0.1×
- Gradual: Jones & Tonetti
as drawn on this page
0.2×
- Conditional: Trammell & Korinek
as drawn on this page
0.6×
- Explosive: Erdil & Besiroglu
as drawn on this page
2.2×
An illustration, not a like-for-like comparison: the IT boom and Goldman Sachs are labor productivity, Acemoglu is total factor productivity, and the four views are growth paths this site drew from each author's headline claim. Bars stop at 10×.
What it means for you
Human bottlenecks move last.
Even in the aggressive scenario, the last tasks to automate are the links with accountability, trust, physical presence, or judgment. Pick a role or business and see where this model puts those links on the timeline.
When does AI stop needing you?
Routine work goes first. What keeps you in the loop is the link AI is worst at: owning the call.
Needed until
~2046
That's about 20 more years of being the weak link AI can't replace.
of the routine work to close the books and sign off on the numbers.
What keeps you needed
- Edge-case judgment
- Audit accountability
- Client trust
The routine ledger work goes first. Putting your name on the numbers, and answering for them, stays human longer than the spreadsheets do.
Live data · Accountants and Auditors: 4 of 30 tasks show AI use, 42% of it automation · 1,449,500 US jobs · median $83,680. See every task →A thought experiment, not a forecast. The years are round anchors on the aggressive scenario: the model's fast case matures around 2050, and links such as accountability, care, physical presence, and the human handshake are placed later, stretching toward 2060. Some may never fully disappear, which is the “if ever” in the question.
Source: Horizon years are site-created illustrative translations of the weak-link automation model, not Jones/Tonetti job forecasts. · Jones & Tonetti (2026), “Past Automation and Future A.I.” · “A.I. and Our Economic Future,” Professor Chad Jones (Stanford GSB)
Note: a “safe until” year is when the remaining human bottlenecks start to matter in the model, not a cliff where the role vanishes overnight. Automation erodes tasks gradually, and can even grow a field: there are more radiologists now than in 2016, not fewer. These are round, illustrative anchors, not job forecasts. Some links, accountability, care, physical presence, may never fully disappear.
The inputs will move.
I track capabilities, adoption, and the bottlenecks that still do not move much. Follow along on LinkedIn.
The fast downside
The downside doesn't wait for the upside.
The same weak-link structure that delays the benefits can make some risks arrive sooner. Strengthening a chain is slow, link by link. Breaking one link is faster.
The bad actor with a jailbroken oracle
Near-termFrontier models keep getting jailbroken, often within days of release. Hand a bad actor a model that can do what the smartest humans can, and ask it to design a pathogen more lethal than Ebola with a three-month latent period. We survived nuclear weapons because only a handful of people held the button. What happens when billions do?
Open-source bug-hunters turned on the grid
1–3 yearsModels are already finding bugs humans missed in decades-old, battle-tested software. It is not hard to imagine a capable open-source version in many more hands within a year or two. How sure are we no one points it at the electric grid, the banking system, or a bio lab? Not an existential problem, but a plausible one, soon.
Retaining power over smarter entities
SpeculativeWhen more advanced species have met less advanced ones in history, it hasn't gone well for the less advanced. Stuart Russell's question is the uncomfortable one: how do we retain power over entities more capable than us, forever?
Slow to improve, fast to fail
StructuralThe same weak-link structure that makes the upside slow makes the downside fast. A chain takes enormous effort to strengthen link by link, but breaking a single link destroys the value instantly. The Space Shuttle Challenger was lost to a $25 O-ring. That asymmetry is the core warning.
Source: The “one broken link destroys the value” fragility draws on Kremer’s (1993) O-Ring theory; the slow-to-strengthen, fast-to-break asymmetry follows the weak-link view in Jones & Tonetti (2026). · Kremer (1993), O-Ring Theory · Jones & Tonetti (2026), “Past Automation and Future A.I.”
Strongest objections
The best arguments against this view.
This page is useful only if its assumptions are visible. These are the objections I think matter most, with responses and what would change the forecast.
This is the crux, and it lands. The forecast's slowness comes from calibrating to a past where the hard tasks never moved much. If AI keeps climbing at exactly the work we've labeled “human,” the gradual story collapses and the upside arrives far sooner. The honest position: the speed depends entirely on how strong those weak links really are, which is the one number we're least sure of.
The gradualism is an assumption, not a law. The cognitive weak links may not hold.
What remains scarce?
Scarce links capture value.
When one input becomes abundant, value often moves to the inputs that remain scarce: judgment, accountability, trust, physical presence, and capital ownership. This is the practical implication of the model.
Be the manager
Someone still decides which output to trust and what tradeoff to make. That judgment matters more when the cost of analysis falls.
Own the capital
If capital captures more of the gains, broad ownership matters more. A slice of the market is one way households share in that shift.
Redistribute deliberately
Abundance makes good outcomes possible, not automatic. Tax-and-transfer is a policy choice, not a law of the model.
Protect non-work meaning
If some work becomes optional, status and purpose still have to come from somewhere: craft, community, family, learning, and taste.
For change leaders: the same story at company scale
The weak links that bottleneck a whole economy have the same shape as weak links inside an organization: handoffs, identity, trust, incentives, and accountability.
| Macro story | Organizational analog |
|---|---|
| Normal tech vs FOOM | Theory of change vs theory of changing |
| Continuation vs break | First-order vs second-order change |
| Weak links bottleneck gains | Sensemaking, identity, coordination cost, OCC |
| Diffusion takes decades | Tsoukas & Chia: organizational becoming |
| Fragile downside | Performing → reinforcing → breakdown → reflecting |
| Scarce links earn returns | Resistance as signal, not obstacle |
The practical move: sequence your automation across the process chain and treat employee resistance as a signal that locates your weakest link, not as friction to overcome.
The library
The research behind the model.
Every claim here traces back to published work. The core weak-link framing and forecast come from Charles I. Jones and Jones & Tonetti; Daron Acemoglu's estimates anchor the skeptical end; capability, compute, and field studies round it out.
Open the research library