Kurzweil Scorecard: The Cooperation Arrived. So Did the Exit.

In 2005, Ray Kurzweil looked at a coronavirus and drew an optimistic lesson. The
SARS outbreak of 2003 had been beaten back in under four months. He read that as
proof of concept: instantaneous global communication had let the world act as a
single organism against a new pathogen, and it would keep doing so. He wrote that
the global response to SARS had been relatively successful and prevented the
outbreak from becoming a true catastrophe, and that international cooperation
would become increasingly vital in confronting future GNR-related challenges,
with organizations such as the World Health Organization needing to be
strengthened (The Singularity Is Near, ch. “Promise and Peril of GNR”).

Twenty years later we have the sequel he was implicitly forecasting — a faster,
deadlier coronavirus — and the results are stranger than a simple hit or miss.
The technical machinery of cooperation performed beyond anything Kurzweil
imagined. The institutional machinery got formalized into the second binding
health treaty in the WHO’s history — and then its largest funder walked out the
same year the ink dried. Kurzweil bet on cooperation. He got cooperation and
its opposite, simultaneously.

The predictions

Three claims sit in this batch, all from Chapter 8, all written when SARS was
the freshest example of a global biological scare. The first is a backward-
looking verdict: SARS was contained, and the world dodged a bullet. The second
is mechanistic: instantaneous global communication enabled a coordinated
worldwide response that would not have been possible in earlier eras. The
third is the forward bet: cooperation becomes more vital over time, and the
WHO gets stronger. Two of these are now testable against a far larger event
than SARS. The third is testable against a treaty signed in May 2025.

Where we actually are

SARS was contained — and that is exactly what made us complacent.

The numbers back Kurzweil cleanly. SARS produced 8,437 probable cases and 813
deaths across 29 countries, a case-fatality ratio near 8.6 percent, and the WHO
declared it contained on July 5, 2003 — less than four months after the global
alert. No vaccine was ever needed; quarantine, contact tracing, and the WHO’s
Global Outbreak Alert and Response Network, then a coalition of 115 health
services and institutions, were enough. On the letter of the claim, Kurzweil is
right. SARS was a genuine success, and it was a cooperative one.

But there is a detail the 2005 framing skips: SARS was containable because it
spread slowly and its carriers were usually visibly sick before they were
infectious. Contact tracing can outrun a pathogen like that. It cannot outrun
one that transmits before symptoms appear — which is precisely the pathogen the
world met in 2020. The success of 2003 was real, and it was also a lucky
match-up. The tools that worked against SARS were quietly declared sufficient,
and the field went quiet with them. Papers on pandemic preparedness sat flat at
roughly 100 to 200 per year from 2005 through 2019. Then, in 2020, the annual
count jumped past 1,900, and it has stayed above 2,000 every year since,
reaching more than 2,600 in 2025 — a tenfold-plus step change that arrived only
after the catastrophe Kurzweil thought SARS had taught us to prevent. Attention
did not accumulate. It waited for the fire.

The communication mechanism didn’t just hold. It broke its own speed record.

Here Kurzweil undersold his own case. He argued that fast global communication
made a coordinated SARS response possible. What happened with COVID-19 is a
version of that mechanism running at a speed the 2005 book would have filed
under science fiction. Chinese authorities released the viral genome on January
11, 2020. Two days later, on January 13, scientists at Moderna and the NIH had
finalized the sequence of an mRNA vaccine against it. The first dose went into a
trial participant’s arm roughly 63 days after that sequence selection. Before
the pandemic, vaccines took five to ten years. In his 2024 follow-up, Kurzweil
himself marveled at the timeline, calling the record-speed vaccine “by far the
most important application of AI to medicine in 2020” and noting it “surely
saved millions of lives.”

The scaffolding underneath that sprint is visible in the record. Patents
mentioning mRNA and vaccines were a trickle through the 2000s; the annual count
reached 9 in 2020 and has climbed to 26 in 2025, with 14 more already in the
first half of 2026. The detection side matured just as fast. US 12,522,863,
granted in January 2026, describes a CRISPR-based cartridge that takes a crude,
unprocessed sample — no lab isolation step — runs an isothermal amplification,
and reads out coronavirus presence on a multiplexed lateral-flow strip. In plain
terms: a field-deployable molecular test that needs no cold chain and no
technician, the descendant of the Cas13 detection chemistry that a 2020 Nature
paper (777 citations) showed could screen for pathogens in massively parallel
fashion. Genomic surveillance, the discipline that tracks a pathogen’s mutations
in near-real time, grew from 27 papers in 2003 to 484 in 2019 to 2,439 in 2025.
Its payoff is concrete: a 2021 Science study (1,527 citations) reconstructed
the emergence and spread of the P.1 variant in Manaus, Brazil, from genomic data
alone, and a 2022 Nature paper (498 citations) caught cryptic variant
transmission in city wastewater before clinics did. This is the coordinated
worldwide response Kurzweil described — sensing, sequencing, and designing
countermeasures across borders in days. On mechanism, reality ran ahead of the
schedule.

Cooperation became vital, got a treaty, and fractured in the same breath.

The third prediction is where the story turns. Kurzweil said cooperation would
grow more vital and the WHO would be strengthened. On paper, both happened. On
May 20, 2025, the World Health Assembly adopted the WHO Pandemic Agreement by a
vote of 124 to none, with 11 abstentions — the product of more than three years
of negotiation launched in COVID’s wake, and only the second binding treaty ever
concluded under Article 19 of the WHO Constitution. The first, fittingly, was
the 2003 tobacco-control convention, adopted the same year as SARS. The 2024
amendments to the International Health Regulations arrived alongside it. If you
score the prediction by whether the institutions of global health response got
formally reinforced, Kurzweil nailed it.

Except the same year told the opposite story. In January 2025 the United States
— the WHO’s largest single funder, contributing roughly $1.28 billion in
2022–23 — began its withdrawal, and it formally exited on January 22, 2026, with
all funding terminated. China has pledged to raise its voluntary contributions
to around $500 million over five years and now stands to become the WHO’s
largest national backer. The treaty strengthening the institution and the
departure hollowing out its budget are not sequential events. They are the same
event, viewed from two capitals. Cooperation did become more vital — vital
enough that its terms are now fought over rather than assumed. The 4,654-citation
Oxford COVID-19 Government Response Tracker, the most-cited pandemic-policy paper
in our records, exists precisely because the response was a patchwork of
national policies that had to be catalogued country by country, not a single
coordinated program. The mechanism Kurzweil predicted is real. It just doesn’t
run through one strengthened WHO. It runs through a contested, multipolar tangle
of treaties, defections, and companies — one of which, the assignee on US
12,412,041 for a privacy-preserving pandemic surveillance system that scores
geographic infection risk from anonymized movement data, is named, with no irony
detectable, Pandemic Insights, Inc.

The scorecard

What Kurzweil missed (and what he nailed)

The pattern in this batch is one Kurzweil’s forecasts show repeatedly:
technology moves faster than he guessed, institutions move messier. He was
right that fast information flow would let the world sense and answer a novel
pathogen — and the two-day genome-to-vaccine sprint of 2020 outran even his
optimism. He was right that cooperation would grow more important. What he
did not model is that importance and agreement are different things. A
capability becoming vital does not make it consensual; it makes it contested.
The more a shared institution matters, the more its funding, its authority, and
its access rules become objects of national struggle. Kurzweil pictured
cooperation as a rising tide. What arrived was a rising stake — higher, sharper,
and pulling in opposite directions at once.

There is a quieter miss underneath the loud one. SARS’s containment was the
evidence Kurzweil built his optimism on, and the containment was partly luck of
the pathogen’s biology. Reading a lucky win as a repeatable capability is the
oldest error in forecasting, and the flat line of preparedness research through
the 2010s suggests the whole field made it, not just one author. The tools that
saved millions in 2020 existed because a handful of labs kept building through
the quiet years. The institutions meant to coordinate them are still arguing
over who pays. Kurzweil bet the machinery would improve and the cooperation
would deepen. The machinery improved faster than he thought. The cooperation
deepened and cracked at the same time.

Method note

This scorecard draws on a corpus of roughly 9.3 million patents and hundreds of
millions of scientific papers, queried for pandemic surveillance, genomic
epidemiology, mRNA vaccine, and pathogen-detection work and counted by year to
establish trends, then read individually for the specific inventions and
findings cited above. Public reporting on the 2003 SARS outbreak, the 2020
COVID-19 vaccine timeline, the 2025 WHO Pandemic Agreement, and the US
withdrawal from the WHO was used to anchor dates and figures. Every number here
traces to a database query or a named source consulted for this post. Kurzweil’s
predictions are quoted from The Singularity Is Near (2005) and his 2024
restatements from The Singularity Is Nearer.