Kurzweil Scorecard: Human Body 2.0

In November 2025, a surgeon at NYU Langone sewed a ten-gene-edited pig
kidney into a patient with end-stage renal disease — the first transplant
in a regulated clinical trial (NCT06878560) sponsored by United Therapeutics.
Nine months earlier, a man in New South Wales had walked out of hospital
with a titanium biventricular pump where his heart used to be, and stayed
out for 104 days until a donor became available. Neither story fits neatly
inside anything Ray Kurzweil wrote in 2005. Both are recognizably his
predictions arriving, refracted.

This batch of seven predictions, from the “on the Human Body” and “on Human
Longevity” chapters of The Singularity Is Near, forecasts that by the
2020s to 2030s we would be printing replacement organs, eating whatever we
wanted while nanobots maintained optimal nutrition, and patrolling our
bloodstream with billions of diamondoid robots. In The Singularity Is
Nearer (2024) he doubled down: “In the 2030s we will reach the third
bridge of radical life extension: medical nanorobots with the ability to
intelligently conduct cellular-level maintenance and repair throughout our
bodies.”

When Kurzweil predicted an outcome, he is usually right or ahead. When
he predicted a mechanism, he is almost always wrong.

Replacement parts: ahead of schedule

Kurzweil in 2005 observed that “technology already provides replacement
parts for hips, knees, shoulders, elbows, wrists, jaws, teeth, skin,
arteries, veins, heart valves, limbs, digits, and increasingly complex
organs such as hearts” (ch. “on the Human Body”). Twenty-one years later,
the definition of “increasingly complex” has moved.

The BiVACOR Total Artificial Heart — a titanium biventricular rotary pump
with a single moving part, a magnetically levitated rotor — bridged all
five patients in its FDA Early Feasibility Study to heart transplants between
July and November 2024. Patient six, in Sydney, was discharged from hospital
with the device implanted and lived at home for 104 days before receiving
a donor heart. BiVACOR is targeting PMA submission for bridge-to-transplant
by late 2027 and permanent implantation by 2029. Meanwhile, bioprinting
patent grants have grown from 3 per year in 2015 to 32 in 2024. Among the
recent filings: US 12,173,310, which claims a hypothermic 3D bioprinting
method for living tissues supported by “perfusable vasculature” — a step
toward the oxygenation problem that has killed earlier printed organs before
they could be implanted. US 12,252,712 describes bioprinted liver-tissue
constructs engineered to exhibit specific liver disorders for drug testing.

And then there is xenotransplantation. Our literature index surfaces a 2022
paper in NEJM on pig-to-human cardiac xenotransplantation (621 citations),
companion work on two pig-to-human kidney xenotransplants the same year
(435 citations), and a 2025 paper on porcine kidney xenotransplantation for
end-stage renal disease (92 citations already). The associated patent
pipeline is thin — only four relevant xenotransplant patents since 2022 —
but the clinical program is real. Kurzweil’s replacement-parts claim is
not only on track; it is accelerating.

Therapeutic cloning: wrong mechanism

Kurzweil predicted by the 2020s that “therapeutic cloning of rejuvenated,
telomere-extended and DNA-corrected versions of a person’s own cells,
tissues, and organs will become part of anti-aging biotechnology”
(ch. “on Human Longevity”). In the Dolly-the-sheep era, cloning was the
obvious path to personalized tissue.

That path closed. The actual entrant to human trials, in January 2026,
was not cloning. Life Biosciences received FDA clearance for ER-100,
delivering three Yamanaka transcription factors (OCT4, SOX2, KLF4) via
intravitreal injection for non-arteritic anterior ischemic optic neuropathy
and open-angle glaucoma (NCT07290244). Altos Labs, sitting on $3 billion,
has not yet named a clinical candidate, but published a 2024 Science
Translational Medicine paper showing targeted partial reprogramming
extends lifespan in mice — with an asterisk: it also raises tumor risk.
The highest-cited literature — a 217-citation 2022 paper on in-vivo
partial reprogramming, a 122-citation 2022 paper on “multi-omic
rejuvenation by a single cycle of transient reprogramming” — points
toward partial reprogramming, not cloning, as the platform. Telomere
extension patents have stayed flat at 14-37 per year since 2015. Kurzweil
got the timing roughly right but bet on the wrong horse.

Personalized nutrition: behind schedule, wrong mechanism

The 2005 prediction was audacious: “ultimately it will be possible to
determine the precise nutrients, including hundreds of phytochemicals,
needed for each individual’s optimal health, and these will be freely and
inexpensively available” (ch. “on the Human Body”). Two decades later,
we do not have a phytochemical prescription. We have ZOE.

ZOE’s 18-week randomized controlled trial with 347 participants, published
in Nature Medicine, showed that a personalized program using postprandial
glucose, triglyceride response, microbiome profile, and health history
beat the standard USDA diet on cardiometabolic and gut-health endpoints.
The 2025 ZOE Microbiome Health Ranking, published in Nature with 34,000
microbiomes analyzed, found that healthy individuals carry an average of
3.6 more “good” microbes than those with diagnosed conditions. US 11,978,543
claims a system for converting microbiome profiling into “microbiome-based
action components” for infants and adults. US 11,754,542 claims a broader
nutritional monitoring and management system.

The mechanism is empirical, not precise. Nobody is computing Kurzweil’s
hundreds of phytochemicals from first principles. Instead, it is black-box
ranking: feed enough postprandial glucose curves and stool samples to a
model, output a food score. The direction is right. The precision he
predicted is not.

Biochemical synthesis: on track, arguably ahead

Kurzweil wrote that within one to two decades of 2005, “humans will be
able to routinely create the vast majority of biochemically relevant
substances” (ch. “on the Human Body”). This one is landing. mRNA-vaccine
patent grants rose from 2 in 2015 to 22 in 2025. AlphaDesign, published
in Molecular Systems Biology in 2025, combines AlphaFold with
autoregressive diffusion to generate de novo proteins, and the authors
demonstrated in-vivo active inhibitors of bacterial phage defense systems.
A 2025 Baker-lab Science paper described modular proteins that can be
designed to bind virtually any DNA sequence. On the small-molecule side,
203 GLP-1-class patents since 2018 and 852 related clinical trials
describe a pipeline that did not exist when Kurzweil wrote. Tirzepatide
alone generated $10.1 billion in a single quarter in 2025. If “routinely
create” means “industrial pipelines exist for arbitrary peptide, protein,
and nucleic-acid therapeutics,” we are there.

Eat whatever you want: wrong mechanism, arriving anyway

The prediction: “at that stage, people will be able to eat whatever they
want for pleasure while maintaining an optimal nutrient flow to the
bloodstream” (ch. “on the Human Body”). Kurzweil imagined this arriving
via metabolic nanobots, absorbing optimal nutrients while the gut enjoyed
whatever it wanted.

What arrived instead was a peptide that makes you want less food. GLP-1
and GLP-1/GIP agonists — Wegovy, Zepbound, Mounjaro — now generate
the largest pharmaceutical revenue streams in history. The phenotypic
outcome Kurzweil described, weight regulation decoupled from moment-to-moment
culinary choice, is partly here. The mechanism is the opposite of what he
predicted: appetite suppression, not nutrient rescue. Dieters on these
drugs do not eat whatever they want. They eat much less of whatever they
want, and a large fraction of them report alcohol and processed food
losing appeal.

Medical nanobots: behind schedule, probably behind the 2030s timeline too

The single 2005 prediction in this batch that has seen essentially zero
real-world progress: “billions of nanobots will travel through the
bloodstream to destroy pathogens, correct DNA errors, eliminate toxins,
and perform many other health-enhancing tasks” (ch. “The Impact”). Our
patent index surfaces twelve relevant “medical nanorobot” patents since
2010. Only one actually claims an embodied device: US 11,324,451, filed
by an independent inventor, describes “medical nanobots with embedded
biosensors for real-time and continuous in-vivo anatomic localization,
diagnosis, disease surveillance, and therapeutic intervention.” It is a
concept patent, not a product. DNA-origami therapeutic filings sit at 18
total. The most-cited preclinical work is a thrombin-delivering DNA
nanorobot that induces tumor-vessel thrombosis in mice and Bama
miniature pigs — impressive, unapproved, and manufactured by hand.
Human trials: zero.

The respirocyte concept — Robert Freitas’s artificial red blood cell that
would let someone hold their breath for four hours — is exactly where it
was in 2005. Perfluorocarbon-based oxygen carriers remain unapproved by
FDA or EMA. The closest thing to a “Human Body 2.0” oxygen-reserve result
in 2025 was a first-in-human safety trial of rectal oxygen delivery
published in Med in October. This is not what Kurzweil had in mind.

The scorecard

What Kurzweil missed, and what he nailed

One pattern dominates this batch: he called the what better than the
how. Replacement parts, rejuvenation biology, on-demand biochemistry,
weight regulation decoupled from willpower — all of it is arriving on
something close to the 2005 timeline. But nanobots did not get us here.
Gene-edited pig organs, titanium rotary pumps, epigenetic reprogramming
proteins, and appetite-suppressing peptides did.

Kurzweil’s 2005 mental model was that biology would be subsumed by
nanotechnology — the body remade as mechanical substrate. The actual path
was that biology got reprogrammed with biology’s own tools: CRISPR editing
of xenotransplant donors, Yamanaka factors to partially rewind cell age,
designed proteins that bind and activate. The substrate stayed wet. The
engineering language is biochemistry, not diamondoid. His “third bridge”
in The Singularity Is Nearer still forecasts diamondoid nanorobots in
the 2030s. On current trajectory, that looks generous by a decade.

Directional calls about outcomes travel well across twenty years.
Specific mechanism calls rarely do.

Method note

This scorecard cross-references Kurzweil’s original predictions against
a patent corpus of 9.3 million US documents, a scientific-literature
corpus of 357 million works, and 541,000 ClinicalTrials.gov studies.
Counts were verified by full-text search; specific patents were read
before being cited. Web searches confirmed trial status, regulatory
milestones, and product timelines from the last twelve months.