Kurzweil Scorecard: The Skies Stayed Quiet. The Black Holes Spoke.

The strangest chapter of The Singularity Is Near is the one that closes the book. Ray Kurzweil ends his treatise on the human-machine merger by stretching past Earth, past the solar system, past biology entirely, to a 22nd-century civilization whose primary preoccupation is how to outrun light. He hangs the prediction on three hedging citations — John Webb’s quasar spectra, Anders Sandberg’s wormhole math, and a Hawking-Preskill bet over a baseball encyclopedia.

Two of those three citations have aged badly. The third grew into one of the most active research programs in theoretical physics — and confirmed Kurzweil’s footnote in a way he could not have foreseen.

The predictions

The chapter — “On the Intelligent Destiny of the Cosmos: Why We Are Probably Alone in the Universe” — bundles a dozen claims, almost all marked “long-term” in our records. Together they form a single argument: intelligence is the universe’s central activity; the speed of light is a soft ceiling, not a hard one; the silence overhead is evidence we are the leading edge of cosmic complexity. The argument leans on four testable empirical hooks Kurzweil cited in 2005: Hawking’s 2004 concession on black-hole information, Webb’s reported variation in the fine-structure constant, Sandberg’s 10⁶⁹-bits-per-second nanowormhole estimate, and the holographic bound’s roughly 10¹²⁰-bit cap on universe content.

Where each hook stands in May 2026 is the actual story.

Where we actually are

The black-hole bet became a research program.

Kurzweil wrote that the Hawking-Preskill bet “was resolved in 2004 when Hawking conceded that information sent into a black hole is not lost and can in principle be recovered” (ch. “On the Intelligent Destiny of the Cosmos”). At the time this was a curiosity — a famous physicist eating his hat in Dublin and handing Preskill a copy of Total Baseball.

In late 2019 and through 2020, Geoff Penington at Berkeley and a parallel team led by Ahmed Almheiri at the Institute for Advanced Study published a sequence of papers reproducing the so-called Page curve — the entropy curve that information has to follow if it really does escape an evaporating black hole — from inside general relativity itself. Almheiri, Hartman, Maldacena, Shaghoulian and Tajdini’s “Replica Wormholes and the Entropy of Hawking Radiation” and Penington’s “Entanglement Wedge Reconstruction and the Information Paradox” sit at 958 and several hundred citations respectively in our local literature index. Almheiri-Engelhardt-Marolf-Maxfield’s “The Page Curve of Hawking Radiation from Semiclassical Geometry” has 735.

The mechanism that ferries information back out is not a quantum trick at the horizon. It’s a geometric structure — an “island” of the black hole’s interior that gets glued, by a quantum path integral, to the outgoing radiation. The interior is, in a precise mathematical sense, partly outside. Hawking lost the bet not because his original calculation was wrong but because it didn’t include enough geometry. The replica-wormhole literature went from six papers in 2019 to thirty-three in 2021 and is still running thirty-plus a year. Quanta Magazine called it “the most famous paradox in physics nearing its end.”

That is more progress on the question than Kurzweil’s 2005 footnote had any right to predict. Verdict: ahead of schedule.

Webb’s quasar lines didn’t replicate.

The bolder hook — “if small experimentally observed variations in the fine-structure constant and speed of light are confirmed, future superintelligence may be able to engineer larger changes” — leaned on John Webb’s 2001 Keck result, extended in 2011 with VLT data, that α had drifted by roughly 10 parts per million across cosmological distances, possibly with a directional pattern across the sky.

The Subaru telescope’s 2017 measurement set tighter limits and found no clean spatial pattern. In 2021–2022 the ESPRESSO spectrograph at the VLT — built with laser frequency comb calibration to remove the systematic errors that contaminated earlier slit-based work — measured Δα/α toward the bright quasar HE 0515−4414 and reported 1.3 ± 1.3 (statistical) ± 0.4 (systematic) parts per million. Consistent with zero. The ESPRESSO team wrote plainly that “on balance, there is currently no compelling evidence for variations in α over cosmological time or distance scales,” and that the earlier dipole was likely a calibration artifact in HIRES and UVES.

Kurzweil’s prediction was hedged with an “if.” The “if” has cashed out unfavorably. Verdict: premise falsified.

The wormhole turned into a paper fight.

Kurzweil’s wormhole citation — “civilization may be able to circumvent the apparent speed-of-light limit, possibly by taking shortcuts through wormholes” (ch. “The Impact”) — leaned on Anders Sandberg’s back-of-envelope that a one-nanometer Einstein-Rosen bridge could push 10⁶⁹ bits per second through itself, given exotic matter and a few engineering miracles.

In late 2022, Daniel Jafferis and a Caltech/Fermilab/Google team published in Nature what was widely covered as “a wormhole on a quantum computer” — a sparsified SYK model run on Google’s Sycamore processor, showing teleportation behavior consistent with the gravitational dual of a traversable wormhole. Bryce Kobrin, Thomas Schuster, and Norman Yao at Berkeley pushed back hard. In 2025, Nature published their formal comment alongside the Jafferis team’s reply. Kobrin et al.’s objection was sharp: the Hamiltonian doesn’t thermalize, the teleportation signal doesn’t resemble traversable-wormhole behavior, and the “perfect size winding” the team reported is an artifact of the Hamiltonian being fully commuting — exactly the property that prevents thermalization.

Twenty years after Kurzweil pointed at wormholes, our most-cited experimental claim is a literature dispute. The gap between “thought experiment” and “working hardware” hasn’t narrowed by a single order of magnitude. Verdict: remains theoretical; the lab demonstration is contested.

The skies stayed quiet.

The chapter’s softest prediction — that “because the skies are quiet despite the expected rapid rise of advanced civilizations, it is likely though not certain that humanity is in the lead” — has aged the best of any in the batch.

In April 2019 the Breakthrough Listen survey, using the Parkes ‘Murriyang’ radio telescope, picked up a narrowband signal at 982 MHz from the direction of Proxima Centauri. They named it BLC1. For eighteen months it was the most interesting candidate technosignature anyone had logged. In late 2021 the team concluded in Nature Astronomy that BLC1 was an “electronically drifting intermodulation product of local, time-varying interferers.” Human equipment. The signal has not recurred in 39 hours of follow-up.

The most exciting current biosignature candidate is a 3-sigma JWST detection of dimethyl sulfide and dimethyl disulfide in the atmosphere of K2-18b, reported by Nikku Madhusudhan’s group at Cambridge in April 2025. A subsequent reanalysis using JWST MIRI data argued there is no evidence of DMS or DMDS at all; laboratory work in 2024 had already shown abiotic DMS formation under simulated exoplanet conditions, and DMS has been detected in cometary environments far from any biology. Madhusudhan called the signal “the strongest hints yet” — but it is not an announcement of life.

Twenty-one years after Kurzweil wrote the chapter, the closest thing to a positive SETI detection is electronic interference, and the closest thing to a biosignature is a contested 3-sigma fingerprint. The Fermi quietness Kurzweil pointed to has, if anything, gotten quieter. Verdict: on track.

The 10¹²⁰ bits held up.

The information-content-of-the-universe claim — “the holographic-universe theory implies a maximum information content for the universe on the order of 10¹²⁰ bits” — has survived as the standard Bekenstein/de-Sitter horizon estimate. Nobody serious is contesting the order of magnitude. Verdict: verified.

The scorecard

What Kurzweil missed (and what he nailed)

Kurzweil’s strongest predictions in the rest of the book are the ones with named years attached. Here, in the chapter that closes the book, almost nothing has a year. He was placing bets on which 2005 footnotes would, by 2026, look prescient. He bet on three.

He won decisively on one: black-hole information escapes, and the way it escapes is one of the central stories of 21st-century physics. The Page curve work pulls quantum gravity, holography, and information theory into a single program. Kurzweil treated the Hawking concession as a one-line aside. It turned out to be a fuse.

He lost the other two. Webb’s variable α was a serious experimental claim in 2005; by 2022 ESPRESSO’s laser-comb calibration had taken it apart. The wormhole speculation became a contested Nature paper twenty years later, no closer to engineering. The forecasting lesson is the asymmetry: the prediction that depended on theoretical structure already in the field — the holographic principle, the Bekenstein bound, the Hawking-Preskill bet — has aged into a richer version of itself. The predictions that depended on specific experimental claims at the threshold of detectability — Webb’s α dipole, Jafferis-style lab wormholes — have aged into disputes.

And then there’s the silence. The skies have stayed quiet through Breakthrough Listen, through TESS, through the JWST. The Fermi argument — that the lack of obvious intelligence overhead is evidence we are early — is the part most easily mocked when you say it out loud, and the part that has gotten quietly stronger every year since 2005. The most rigorous candidate signal of the last decade was a Parkes-area oscillator. The most rigorous biosignature is a 3-sigma fingerprint that may not survive a single MIRI reanalysis.

If Kurzweil’s overall thesis — that intelligence is the central activity of the universe — has any teeth, it now leans much harder on the silence than it did in 2005. The black-hole story gives the thesis its physics. The Webb story takes away a piece of its engineering. The SETI story, awkwardly, gives it its strongest empirical support: nobody else seems to be doing the thing yet.

Method note

Literature counts came from our local OpenAlex-derived index of roughly 357 million papers; trends on the Page curve, replica wormholes, fine-structure-constant variation, and Fermi-paradox literature were pulled from year-by-year tallies on full-text-indexed abstracts. Citation counts on Penington and Almheiri were checked against the JHEP records. Specific findings — the ESPRESSO 1.3 ppm result, the Kobrin-Schuster-Yao comment and its 2025 Nature reply, the BLC1 interference conclusion, the K2-18b DMS reanalysis — came from web searches against the primary papers and reporting from Quanta, NPR, and the Nature news team. Predictions were sourced from the chapters “On the Intelligent Destiny of the Cosmos” and “The Impact” in The Singularity Is Near (2005); the cosmos appendix is one of the things Kurzweil did not significantly revisit in The Singularity Is Nearer (2024).