Kurzweil Scorecard: Bridge One and the HDL Gamble

In 2005, Ray Kurzweil made a specific, falsifiable wager about how the first wave of “radical life extension” would arrive. His two exhibits in The Singularity Is Near: a Pfizer cholesterol drug called torcetrapib, and a recombinant protein called Apo-A-I Milano that had shrunk plaque by 4.2 percent in forty-seven patients. Together with “aggressively applied” lifestyle measures, these were supposed to be Bridge One — the bridge that kept baby boomers healthy long enough to reach the AI-and-biotech bridge of the 2020s.

Twenty years later, both exhibits are wreckage. Torcetrapib killed people in its pivotal trial. Apo-A-I Milano was bought, resold, reformulated, and finally laid to rest in 2024 when a 18,200-patient Phase 3 trial of its closest living descendant missed its primary endpoint. And yet — and this is the uncomfortable part — Kurzweil was directionally right. Coronary plaque does now regress in human beings on contemporary therapy. It just happens by mechanisms he did not name in a single chapter of the book.

The predictions

The batch contains four linked predictions about what Kurzweil called “combating heart disease” and Bridge One to longevity:

• Bridge One. “Existing medical and nutritional knowledge, if aggressively applied, is sufficient to dramatically slow aging and reverse major disease processes so that baby boomers can stay healthy long enough to reach later biotechnology advances” (ch. “Designer Baby Boomers”).
• The torcetrapib claim. “Pfizer’s Torcetrapib, a CETP-blocking HDL-boosting drug, was in phase 3 FDA trials in 2005, with Pfizer spending a record \$1 billion and planning to combine it with Lipitor” (ch. “Combating Heart Disease”). A factual claim about 2005.
• The Apo-A-I Milano claim. “Recombinant Apo-A-I Milano caused rapid regression of atherosclerotic plaque in animal trials and reduced plaque by an average of 4.2 percent after five weekly infusions in a 47-subject phase 1 FDA human trial” (ch. “Combating Heart Disease”).
• The directional forecast. “New precisely targeted drugs will reverse atherosclerosis by intervening in key steps of the disease process” (ch. “Combating Heart Disease”), timed for the 2010s.

The first and fourth are predictions. The middle two are factual claims offered as evidence that the fourth was already in motion. How did each fare?

Torcetrapib: the catastrophe

Kurzweil’s facts about 2005 are accurate. Pfizer really was in Phase 3, really had spent about a billion dollars, and really did plan to sell torcetrapib fixed-dose with Lipitor. The trial design was massive: 15,000 subjects with coronary heart disease, randomized to torcetrapib plus atorvastatin or atorvastatin alone.

On December 2, 2006 — roughly a year after The Singularity Is Near was published — Pfizer halted the ILLUMINATE trial. The independent data and safety monitoring board had found 82 deaths in the torcetrapib arm versus 51 in the atorvastatin-only arm. The outcomes trial of a drug designed to prevent heart attacks was killing people faster than the control.

The NEJM report of ILLUMINATE and its two imaging-endpoint companion trials (962, 660, and 455 citations respectively) established the clinical verdict. A second wave of papers pinned down the mechanism: torcetrapib raised blood pressure by three to four millimeters of mercury and induced aldosterone production by an off-target pathway independent of CETP inhibition (Circulation 2008; Endocrinology 2009). The molecule had a hidden pharmacology that the ten-thousand-patient Phase 2 program had failed to surface.

The CETP inhibitor class limped on for another decade. Dalcetrapib (Roche) failed for futility in 2012. Evacetrapib (Lilly) was terminated in 2015. Only anacetrapib (Merck) ever showed benefit, and Merck declined to file it because the effect was too small and the drug accumulated in adipose tissue for years (Clinical Pharmacology & Therapeutics 2017). The 2017 review “Trials and Tribulations of CETP Inhibitors” (Circulation Research, 282 citations) closed the chapter. New CETP-inhibitor patent filings peaked in 2015 in our records and collapsed to single digits after 2018.

Verdict: Behind schedule — catastrophically. The 2005 factual claims were accurate. The implied forecast that torcetrapib would be a cornerstone of atherosclerosis reversal was wrong by the time the book’s paperback shipped.

Apo-A-I Milano: a twenty-year autopsy

The second exhibit has a stranger arc. Apo-A-I Milano is a naturally occurring variant of the main HDL protein, identified in a family from Limone sul Garda who carried sub-normal HDL levels but essentially no coronary disease. Pfizer acquired the developer, Esperion, in 2003 for \$1.3 billion on the strength of the data Kurzweil cited — then shelved the program. A series of successors tried to revive the thesis: CER-001 from Cerenis, MDCO-216 from The Medicines Company, and CSL112 from CSL Behring. Each failed in turn.

The final verdict came in February 2024. The AEGIS-II trial randomized 18,219 patients with recent acute myocardial infarction to four weekly infusions of CSL112 or placebo. At 90 days, the primary composite of cardiovascular death, MI, or stroke occurred in 4.8 percent of the CSL112 arm and 5.2 percent of placebo — hazard ratio 0.93, confidence interval 0.81 to 1.05, not significant. tctmd’s coverage called it “a new blow to the HDL hypothesis.” The most recent patent in our records mentioning Apo-A-I Milano is US 10,391,144, issued in 2019 — an AAV8 gene-transfer approach that reprograms plaque macrophages from an inflammatory to an anti-inflammatory phenotype. A reasonable scientific idea; not a drug anyone is taking.

Verdict: Wrong mechanism. The HDL-raising strategy that Kurzweil treated as imminent turned out to be an elegant hypothesis that repeatedly failed in humans.

What actually reverses plaque

Here is where Kurzweil deserves real credit. He predicted that “precisely targeted drugs will reverse atherosclerosis” by the 2010s. They do. They just don’t work the way he described.

In November 2016, Steven Nissen — the same cardiologist who ran ILLUMINATE a decade earlier — presented GLAGOV at the American Heart Association. The trial used intravascular ultrasound to measure coronary plaque volume in 968 statin-treated patients randomized to evolocumab (a PCSK9-inhibiting antibody) or placebo for eighteen months. Placebo plaque volume increased by 0.05 percent; evolocumab plaque volume decreased by 0.95 percent. Sixty-four percent of treated patients regressed, versus forty-seven percent of controls. Mean LDL cholesterol fell to 36.6 mg/dL on drug versus 93.0 mg/dL on placebo.

That was the mechanism Kurzweil didn’t name. PCSK9 is not discussed in either the 2005 or 2024 book. The patent class took off after 2017; by 2024 we count fourteen granted PCSK9-related atherosclerosis patents, including methods-of-use claims from the sponsors of the two approved antibodies. FOURIER (2017) reported a 15 percent MACE reduction across 27,564 patients on evolocumab plus statin. ODYSSEY OUTCOMES showed similar benefit for alirocumab. Inclisiran, a siRNA silencing PCSK9 messenger RNA with two injections per year, was approved in 2020.

The second unforeseen mechanism is inflammation. Phase 3 atherosclerosis trials starting after 2018 are dominated by drugs the 2005 book did not anticipate: ziltivekimab (IL-6 receptor antagonist), colchicine (generic anti-inflammatory), olezarsen (antisense oligonucleotide targeting ApoC-III), and orforglipron (oral GLP-1 receptor agonist). The SELECT trial of semaglutide reported in 2023 a 20 percent MACE reduction in 17,604 overweight-or-obese patients without diabetes — a mechanism with no clean analogue in The Singularity Is Near.

Verdict for the directional forecast: On track in outcome, wrong in mechanism. The 2010s did produce drugs that demonstrably reverse coronary plaque. Not one of them appears in Kurzweil’s argument.

Bridge One, revisited

The fourth thread is the Bridge One thesis itself: that aggressively applied existing knowledge would dramatically slow aging for people who followed it. This is the thesis Kurzweil softened in 2024. The Singularity Is Nearer concedes: “life expectancy gains in developed countries have slowed since roughly the middle of the twentieth century. For example, from 1980 to 2000…it only increased from seventy-four to seventy-six” (ch. “Why Now?”). The first bridge, in his current framing, can “only delay the inevitable.”

CDC data confirms the slowing. US life expectancy at birth was 77.8 years in 2005, when The Singularity Is Near was published. In 2024, after a COVID-era drop and rebound, it reached 79.0 years — the highest ever recorded, and an increase of 1.2 years across two decades. The cohort born in 1946 (the oldest baby boomers) would now be 80, within the expected lifespan rather than dramatically beyond it. Whatever Bridge One did, it did not visibly change the slope.

Verdict: Behind schedule. The population-level claim did not come true. Kurzweil himself appears to have moved the thesis forward to Bridge Two (AI-plus-biotech, 2020s) and Bridge Three (medical nanorobots, 2030s).

The scorecard

What Kurzweil missed (and what he nailed)

The pattern is sharp. Kurzweil’s directional forecast — targeted drugs would reverse atherosclerosis within fifteen years — came true. His specific mechanisms were zero-for-two. The drugs that actually produced plaque regression (PCSK9 antibodies, PCSK9-targeting siRNA, IL-6 receptor blockade, GLP-1 agonism) received nothing more than a passing mention in the 2005 book.

When The Singularity Is Near was written, HDL-raising was the fashionable hypothesis in cardiology and CETP inhibition was the hottest program in drug development. Kurzweil didn’t invent the bet; he picked it up from the contemporary consensus. The mechanisms that actually worked — monoclonal antibodies against a protease nobody had emphasized, hepatocyte-targeted siRNA, a diabetes drug for non-diabetics — were not fashionable, and in some cases not yet named, in 2005. The book was right about the destination and confident about a road that turned out to be a dead end.

The Bridge One thesis has aged worse. In 2024 Kurzweil writes as if the first bridge’s role was always to buy time rather than deliver a dramatic effect. In 2005 the book implied the effect was within reach — that baby boomers could live long enough to reach the next generation of therapies simply by taking supplements, eating well, and asking their doctors about torcetrapib. The aggressive version of that claim no longer has a defender.

Method note

This scorecard draws on our own database of 9.3 million US patents, 357 million scientific papers, and 541,000 clinical trials, cross-checked against published trial reports, FDA actions, and CDC mortality data. Patent-count trends come from full-text searches on class keywords; claim-level descriptions come from reading the actual claims and abstracts of named patents. Citation counts reflect what we had indexed as of April 2026.

Sources include: Wikipedia — Torcetrapib; NEJM — ILLUMINATE (10.1056/NEJMoa0706628); Nissen et al., JAMA 2016 — GLAGOV; CSL press release — AEGIS-II top-line results; European Heart Journal — AEGIS-II full analysis; Sabatine et al., NEJM 2017 — FOURIER; Lincoff et al., NEJM 2023 — SELECT; CDC Data Brief No. 521 — US life expectancy 2023.