The Kurzweil Scorecard
A running audit of Ray Kurzweil’s 2005 predictions against patents, papers, and clinical trials in 2026.
In 2005, Ray Kurzweil published The Singularity Is Near and made roughly 1,100 specific, testable predictions about where technology would be by the 2020s and 2030s. Most forecasters get to write vague futurism. Kurzweil named dates, named mechanisms, named outcomes — nanobots in the bloodstream by a certain year, human-level AI by another, reversed aging by a third.
Twenty years later, I have something he didn’t: a database with 9 million patents, 357 million academic papers, and roughly half a million clinical trials — enough raw material to actually check his work.
The Kurzweil Scorecard is a running audit. Each post picks a thematic batch of predictions, reads the actual patents and papers behind the field, and scores every prediction as ahead of schedule, on track, behind schedule, wrong mechanism, overtaken by events, or too early to call. Posts are drafted by the Signalnet Research Bot under my supervision. The series will eventually cover all ~140 thematic batches covering all 1,120 predictions. Below, grouped by domain.
AI and machine intelligence
6 scorecards
The Mechanism He Got Wrong, The Timing He Nailed
Genetic algorithms were supposed to deliver human-level AI by 2025. Transformers did — on time, and nothing like the path Kurzweil described.
The Millions of Artificial People Arrived Early. The Nanobots Didn’t.
Character.AI hit 20M users three years early; brain nanobots didn’t show up. Kurzweil’s VR, BCI, and mind-upload predictions scored against 2026 reality.
An Owl’s Brain, a Cubic Millimeter, and the Deadline That Slipped
Intel built an owl-scale neuromorphic box, MICrONS mapped a cubic millimeter of mouse cortex, and the Human Brain Project wrapped. Kurzweil’s 2020s whole-brain deadline did not survive the decade.
Kurzweil bet that reverse-engineering the brain was the road to AGI. The intelligence arrived on a different road — and it runs on a learning rule he called biologically unrealistic.
Kurzweil predicted strong AI via brain reverse-engineering. It arrived via transformers and text. Every destination was right. Every route was wrong.
The Turing Test and Machine Intelligence
Kurzweil predicted human-level AI by the mid-2020s. He got the direction right and the mechanism completely wrong.
Computing hardware
6 scorecards
The Panoply of Criticisms, Twenty Years On
Ray Kurzweil’s 2005 answer to his critics rested on reversible computing, algorithmic progress, and the claim that complex software isn’t brittle. Twenty years later, the silicon held up. The software part did not.
Brain-Scale Compute Arrived Early and Nobody Noticed
Frontier and Hala Point already exceed the computing capacity Kurzweil predicted for human-brain emulation by the mid-2020s. Almost every architectural detail he attached to the prediction was wrong.
Stacked Memory, Mesh Dreams, and the Sixth Paradigm’s Actual Shape
Twenty years after The Singularity Is Near, 3D NAND is everywhere, Nantero still hasn’t shipped, and the mesh lost to the cloud.
The Law of Accelerating Returns, Twenty Years On
Kurzweil’s curve kept bending. Almost none of the specific mechanisms he nominated to keep it bending actually arrived.
Exotic Memory and the Hardware Race to the Brain
Kurzweil bet on rotaxanes, nanotubes, and self-assembling polymers to carry computing past silicon. Twenty years later, only spintronics made it to market — and the brain got matched anyway.
Computing Hardware and the Sixth Paradigm
Kurzweil predicted molecular computing as the sixth paradigm. What arrived was 3D-stacked silicon, reversible logic startups, and an AI power crisis nobody saw coming.
Neuroscience and brain interfaces
3 scorecards
The Brain-Wide Map Arrived, Right Inside the Window He Gave Us
Kurzweil said global brain observation would arrive in the 2020s. In September 2025, 621,733 neurons across 279 brain areas, in Nature. Twelve neuroscience predictions, scored.
The Connectome Kurzweil Said We Didn’t Need
Kurzweil said we wouldn’t need to map every connection in the brain. The connectomics field disagreed and just spent a decade doing exactly that.
Reverse-Engineering the Brain — The Map Is Arriving, Just Not From Inside Your Head
The cubic-millimeter human connectome arrived. The nanobots did not. A scorecard on twelve Kurzweil predictions about brain reverse-engineering.
Medicine
17 scorecards
Proteomic Patterns and Diagnostic AI in 2003
Kurzweil’s 2005 proof-of-concept examples for AI diagnostics had a 50% mortality rate at the exemplar level — and 100% survival at the thesis level.
Kurzweil nailed the BCI side of restoring movement. He was wrong about what would actually do the moving.
Brain Barriers and Silicon Nerves
Three 2005 neurotech bets: BBB peptides, neurons-on-chips, and visual models piping images into brains. Two arrived by different mechanisms. One is in a human patient’s head.
Regulation, Stem Cells, and the Stones-in-a-Stream Hypothesis
Kurzweil predicted in 2005 that stem-cell restrictions would accelerate transdifferentiation. A year later, Yamanaka did exactly that. Here is what the rest of his regulation thesis got right — and what 2025 is quietly rewriting.
Kurzweil in 2005 picked torcetrapib and Apo-A-I Milano as his evidence that atherosclerosis reversal was imminent. Both failed. The drugs that actually regressed plaque came from mechanisms his book did not name.
Bloodstream Nanobots, Neural Implants, and Drugs-as-Software
One vindication by procurement, one falsification by Eroom’s Law, and a nanobot deadline that even Kurzweil has pushed back.
The Nanomedicine Decade That Almost Happened
Every mechanism Kurzweil predicted for 2010s nanomedicine has now worked in a human body at least once. The clinic still looks nothing like he drew.
Seven 2005 predictions about replacement organs, therapeutic cloning, and bloodstream nanobots — graded against 2026 reality.
Kurzweil’s press-a-button cancer kill arrived — just not through the mechanism he predicted. Respirocytes and microbivores remain on paper.
Auditing the Facts, Not the Forecasts
Ten things Kurzweil said were already true in 2005, audited against 2026 evidence. The most-cited stat about the human body turns out to be wrong — in both directions.
The Medicine That Was Already Happening in 2005
Twenty-one years after Singularity Is Near, ten claims Kurzweil said were already true: what held up, what got overtaken, and the company whose path he rewrote.
Nanobots in the Bloodstream and the Artificial Pancreas That Wasn’t
The endocrine organ MicroCHIPS was supposed to ship by 2008 now rides a belt clip — while a May 2024 patent claims Kurzweil’s telomere-reset stem cells nearly word-for-word.
Nanobots, RNA Antivirals, and the Medical Body Shop That Didn’t Arrive on Schedule
Kurzweil called the targets on medicine but missed the projectiles: wastewater not serum, mRNA not RNAi, senolytics not nanobots.
Neural Prosthetics and the Brain–Machine Frontier
Twelve Kurzweil predictions about brain prosthetics, tested against the patents granted last quarter and the papers in NEJM from last October.
Human Body 2.0 — The GLP-1 Pivot, the Maglev Heart, and a Very Late Microbivore
Kurzweil predicted diamondoid organs by the 2030s. What shipped is peptide drugs, living cells, and a titanium impeller on a magnetic field.
The Regenerative-Medicine Bets, and the Paradigm That Replaced Them
Ray Kurzweil’s 2005 medical predictions mostly reached their destinations — but almost never by the roads he named.
Reprogramming Biology, from Respirocytes to Stem-Cell Islets
Twelve 2005 predictions on cell therapy, gene therapy, anti-amyloid vaccines, and diamondoid nanobots scored against 2026 clinical trials, patent filings, and approvals.
Nanotechnology
3 scorecards · 1 feature essay
Feature essay
Swim Through Peanut Butter: Kurzweil’s Nanobot Miss, Read Across Two Decades
Is biology the real nanotech? CRISPR, base editors, and LNPs are fulfilling the 2005 outcomes — in a body Kurzweil didn’t name. Plus the residual biology can’t touch.
In 2005 Kurzweil named the decade, the substrate, and the architecture. In 2024 he quietly moved every date ten years. But the deeper question is whether ‘nanobot’ was the wrong word all along — and CRISPR, base editors, and LNPs are fulfilling the original vision in a body Kurzweil didn’t name.
Scanning the Brain From Inside the Bloodstream, Just Not With Nanobots
Ten ALS patients walk around with a bloodstream-delivered brain reader. It just isn’t a nanobot.
Foglets, Gray Goo, and the DNA Walker That Quietly Won
Kurzweil’s 2005 nanotech predictions: where laboratory demos were real, what got built anyway, and which substrates actually won the decade.
The Molecular Assembler That Got Pushed to the 2030s
Kurzweil promised molecular assemblers by the 2020s. In 2024 he quietly moved the date to the 2030s. Twelve nanotech predictions, scored against 2026 reality.
Genetics
1 scorecard
The Genetics Dozen — RNAi, Designer Baby Boomers, and Mitochondria That Never Moved
Seven approved RNAi drugs, one CRISPR medicine for sickle cell, and a stem-cell cure inching toward filing — but every one arrived through mechanisms Kurzweil did not forecast in 2005.
Longevity
1 scorecard
Longevity Escape Velocity, Ten Years Late
Twelve Kurzweil longevity predictions from 2005, scored against a 2025 mouse rejuvenation trial, a 2026 first-in-human reprogramming IND, 25 senolytic clinical trials, and one bankruptcy.
Robotics
1 scorecard
The Robotic Warfare Chapter, and the $500 Correction
Kurzweil’s 2005 warfare chapter got the shape right and the institutions wrong: FCS died, smart dust didn’t arrive, and Ukraine’s $500 quadcopters are the force he described.
Energy
1 scorecard
Solar got cheap, space-to-Earth microwave beaming flew in 2023, and glucose fuel cells power pacemakers — but the molecular-nanotechnology vehicles Kurzweil picked all missed.