Whose Evolution Is Faster? 5 Powerful Technologies Racing Ahead of Their Human Builders

By Dr. Narayan Rout |Author | Researcher |     lP10: The Next Human — Science, Technology, and the Future We Are Already Building  ·  34 min read  ·  Published: June 19, 2026

Contents In This Research Pillar

Table of Contents

1. Introduction
2. 1. What Does “Evolution Speed” Actually Mean? Comparing Timescales Properly
3. 2. AI — 300,000x Compute Growth Since 2012, and What “3.4 Months” Actually Means
4. 3. CRISPR — From a 2012 Discovery to a 2025 Personalized Gene Therapy
5. 4. Quantum Computing — Google’s Willow Chip and the 10^25-Year Comparison
6. 5. Robotics — Real Progress, Real Setbacks, and a Genuinely Uneven 2025-2026
7. 6. Space Exploration — Reusable Rockets, Real Catches, and a Self-Assigned “50-50” Mars Timeline
8. 7. Why Human Evolution Cannot Keep Pace with Technological Evolution
9. So Whose Evolution Is Actually Faster? The Honest Synthesis
10. The Quest Sage Insight
11. What You Can Do With This
12. Conclusion: Two Clocks, Running at Once
13. Frequently Asked Questions: Technology Growth Rates and Human Evolution
14. References and Sources
15. Further Reading

Here’s a strange thought experiment worth sitting with for a moment. The human genome that built your brain took roughly two million years to acquire its last major adaptive change. The AI model that can now hold a conversation about that very fact didn’t exist three years ago, and the compute power behind it has grown roughly 300,000-fold since 2012. Something is racing forward at a pace evolution never had access to — and it isn’t us, not in the biological sense. It’s what we’ve built.

This article takes that observation seriously enough to actually test it against the evidence, rather than just gesturing at it the way a lot of techno-futurist writing tends to. Five technologies, five very different domains, five distinct growth curves: artificial intelligence, gene editing, quantum computing, humanoid robotics, and space exploration. Each one gets examined on its own terms, with real numbers and named sources, not just vibes about “the pace of change.”

And here’s the thing the honest version of this story requires: not every one of these five curves is smooth. AI and quantum computing really are advancing at a pace that has no real precedent. Humanoid robotics and space exploration are advancing too, but unevenly, with public setbacks and admitted failures sitting right alongside genuine breakthroughs — which is, if anything, a more interesting and more trustworthy story than a uniform tale of unstoppable acceleration. By the end, this article arrives at a specific, evidence-based answer to the question in its title, and it’s a more useful answer than a simple “yes, technology is winning.”

Before comparing technology’s pace to human evolution, it’s worth being precise about what each term actually refers to, because conflating them is where a lot of casual commentary on this topic goes wrong.

Human biological evolution operates through genetic mutation, natural selection, and inheritance across generations — a process inherently bound to reproduction cycles measured in decades, with major adaptive changes in the human lineage typically taking, per standard evolutionary biology timescales, somewhere on the order of hundreds of thousands of years to roughly two million years to become established and widespread. Technological evolution operates through an entirely different mechanism: deliberate human design, rapid iteration, and compounding knowledge transfer that doesn’t wait for anyone to be born or die. That’s the structural reason the speed gap exists at all — it isn’t that humans are evolving slowly out of some failure. It’s that biological evolution and technological development are two different processes running on two different clocks, and only one of those clocks has been freed from the constraint of generational reproduction.

Computer scientist and futurist Ray Kurzweil has argued, in what he calls the law of accelerating returns, that the pace of technological change itself increases exponentially over time — not just within any single technology, but across technology as a category, generalizing the logic of Moore’s Law into a broader claim about innovation itself. This article doesn’t take that claim on faith. It tests it directly against five specific, current technologies, starting with the one currently moving fastest by any measure available.

Artificial intelligence offers the clearest, best-documented case of genuinely unprecedented technological acceleration, and the data behind it comes from serious quantitative research, not marketing copy.

The 2022 paper ‘Compute Trends Across Three Eras of Machine Learning’ by Sevilla and colleagues traced how the computational power used to train AI systems grew, for the first few decades of the field, along a roughly Moore’s-Law-like pattern — doubling approximately every two years, tracking the growth of transistor density in computer chips generally. That changed at the start of what researchers call the Deep Learning era, marked by the 2012 image-recognition system AlexNet, when the doubling time compressed dramatically to roughly six months as researchers began investing far more heavily in dedicated computational power. With the 2015 emergence of large-scale AI systems like AlphaGo, a third era began, and by the mid-2020s, multiple independent analyses describe AI training compute as having grown more than 300,000-fold since 2012, with a doubling time of approximately 3.4 months — an order of magnitude faster than Moore’s Law’s historic two-year cycle.

Researcher Nabeel Qureshi’s ‘human brain equivalents’ (HBE) framework offers a genuinely useful way to make this abstract growth curve concrete. The framework estimates the total computational power added globally each year, expressed in units roughly equivalent to one human brain’s processing capacity. In 1993, the entire world’s added computing power amounted to roughly one HBE. By 2015, that figure had grown to roughly one million HBEs added per year. By 2026, on the same trendline, the estimate runs to roughly one billion HBEs added annually — a billionfold increase in barely three decades, against a human population that, over the same period, grew by less than double. (For the deeper philosophical implications of this gap, see Yogic Intelligence vs Artificial Intelligence, TheQuestSage.com, Sl 67, and Carbon vs Silicon Intelligence : 5 Fundamental Differences, TheQuestSage.com, Sl 68.)

If AI represents the fastest-moving technology examined in this article, CRISPR gene-editing offers the clearest illustration of how quickly a fundamental scientific discovery can translate into real clinical practice once the underlying mechanism is understood.

In 2012, Jennifer Doudna and Emmanuelle Charpentier demonstrated that CRISPR — a defense mechanism bacteria use to recognize and cut viral DNA — could be reprogrammed to target and edit any organism’s DNA, including human DNA, with a precision and ease that earlier gene-editing techniques could not match. The discovery earned both researchers the 2020 Nobel Prize in Chemistry, but the more striking story is what happened in the years between the lab discovery and real clinical application.

By 2023, more than 200 people had undergone experimental CRISPR-based therapies. In May 2025, physicians at Children’s Hospital of Philadelphia (CHOP) used CRISPR-based gene editing to treat a child with a rare genetic disorder through a treatment personalized specifically to that patient’s unique DNA sequence — a meaningfully different tier of medicine from earlier CRISPR treatments, which targeted only well-characterized, common mutations shared across many patients. The industrial and commercial scale of this growth is just as striking: by early 2025, the United States had 217 dedicated gene-editing companies, against a few dozen in Europe and roughly 30 in China, and the global CRISPR market — valued at $3.12 billion in 2022 and $4.69 billion in 2024 — is projected to reach approximately $33 billion by 2033, according to Grand View Research and Statista analysis. From bacterial defense mechanism to personalized human gene therapy and a multi-billion-dollar industry, in barely thirteen years.

Quantum computing spent over a decade being described, fairly, as a field full of theoretical promise but persistent practical obstacles — chiefly, that quantum bits (qubits) are so sensitive to their environment that errors accumulate faster as systems scale up, seemingly blocking the path to genuinely useful quantum computers. December 2024 changed that calculus directly.

Google’s Willow processor, a 105-qubit superconducting quantum chip, demonstrated for the first time that error rates could be suppressed exponentially as more physical qubits were added to form each logical qubit — a long-sought threshold described in a paper published in Nature, and described across the industry as quantum computing finally going ‘below threshold.’ To illustrate the practical implication, Willow completed a standard random-circuit-sampling benchmark calculation in under five minutes; the fastest classical supercomputer in existence would require an estimated 10^25 years to perform the equivalent calculation — a number so large it has no real intuitive meaning, which is itself part of the point.

Through 2025, the field built directly on this result. Google’s ‘Quantum Echoes’ algorithm achieved a verified 13,000-times speedup over classical supercomputers on a molecular-simulation task, described as the first fully verifiable demonstration of real-world quantum advantage. IBM announced a fault-tolerant roadmap centered on a system called Quantum Starling, targeting 200 logical qubits by 2029. And in March 2025, IonQ, partnering with the engineering simulation company Ansys, ran a real medical-device simulation on a 36-qubit quantum computer that outperformed classical high-performance computing by approximately 12% — a comparatively modest number, but significant precisely because it was a genuine practical application, not a specially constructed benchmark designed to favor quantum hardware. The table below situates these three 2025 milestones side by side.

Milestone	Date	What It Demonstrated
Willow below-threshold error correction	December 2024	Error rates fall, not rise, as qubit count scales — the core barrier to practical quantum computing
Quantum Echoes algorithm	2025	13,000x verified speedup over classical supercomputers on a real molecular-simulation task
IonQ + Ansys medical-device simulation	March 2025	~12% real-world advantage over classical HPC — first practical (not benchmark-only) quantum edge

(For a deeper treatment of where this leaves quantum computing’s remaining open problems, see Quantum Computing Explained: 5 Problems, TheQuestSage.com, Sl 69.) The pattern across all three milestones is the same one seen in AI and CRISPR: a technology moving from theoretical promise to demonstrated, verifiable capability on a timescale of months, not decades.

Humanoid robotics is where this article’s pattern of clean exponential acceleration breaks down, and that break is itself an important, honestly-reported finding rather than a gap in the story.

Figure AI has published customer-corroborated, operational data that is hard to dismiss as hype: its Figure 02 and Figure 03 humanoid robots have supported production of more than 30,000 BMW vehicles, with over 1,250 hours of logged operational runtime, and its dedicated BotQ factory reached a production rate of one robot per hour by June 2026. Boston Dynamics’ fully electric Atlas began its first 2026 production deployments to partners including Hyundai and Google DeepMind. Chinese manufacturer Unitree shipped approximately 5,500 humanoid units across 2025 alone, targeting 10,000 to 20,000 for 2026.

Tesla’s Optimus program, despite vastly larger resources, tells a more complicated story, and reporting it honestly matters more than reporting it impressively. On Tesla’s own January 2026 Q4 2025 earnings call, CEO Elon Musk acknowledged that despite earlier public claims of over 1,000 deployed Optimus units, none were actually performing useful work in Tesla’s factories — describing the program candidly as ‘still very much at the early stages’ and ‘still in the R&D phase.’ Independent reporting from Bloomberg, The Verge, and Electrek separately documented teleoperation — a human operator controlling the robot remotely — during several public Optimus demonstrations in 2024 and 2025, a detail not initially disclosed at the events themselves.

And yet, in the same window, a genuinely remarkable autonomous achievement occurred: on April 19, 2026, a fully autonomous humanoid robot named ‘Lightning,’ built by the Honor and Monkey King team, won the Beijing E-Town Half-Marathon in a time of 50 minutes 26 seconds — beating the existing human world record by nearly seven minutes. The honest picture of humanoid robotics in 2025-2026, in other words, is neither pure hype nor pure stagnation. It is uneven, genuinely fast in specific narrow domains, and genuinely behind schedule in others — a more useful and more credible picture than either extreme.

Space exploration follows a similar pattern to robotics: genuine, hard-won technical milestones sitting alongside an honestly hedged, frequently-slipping timeline for the most ambitious goals.

SpaceX’s Starship program achieved a genuine first in the history of rocketry during 2025: its Super Heavy booster, after returning from launch, was caught directly by the launch tower’s mechanical arms — nicknamed ‘chopsticks’ — demonstrating the core mechanism required for the kind of rapid, low-cost reusability that has never previously been achieved at this scale. SpaceX went on to refurbish and relaunch a caught booster within roughly a week, and completed the first in-orbit propellant transfer test, a capability NASA’s Artemis program requires for its Starship-based Human Landing System, which NASA currently targets for a crewed lunar landing in 2027.

The same period included real, public setbacks. Starship test flights in January and March 2025 experienced significant in-flight anomalies, and a May 2025 flight lost vehicle control after an apparent propellant leak, resulting in an uncontrolled reentry over the Indian Ocean. Presenting a detailed development timeline against this backdrop, Elon Musk assigned only a ’50-50 chance’ to achieving an uncrewed Starship landing on Mars by the end of 2026 — the next available low-energy Earth-Mars transfer window, occurring roughly every 26 months due to orbital mechanics — with crewed landings not anticipated before 2028 or 2031 depending on how the 2026 attempt unfolds. If the 2026 window is missed, Musk himself noted the next viable attempt would not come until the following alignment roughly two years later, illustrating how planetary mechanics impose a hard external constraint that no amount of engineering acceleration can simply override.

This is a genuinely useful contrast to keep in mind against the AI and quantum computing sections above: some technological frontiers really do compress timescales dramatically. Others remain bound by physical constraints — orbital mechanics, in this case — that no exponential curve in computing power can shortcut.

Humanity has transformed the world at a breathtaking pace. In just a few centuries, we progressed from horse-drawn carts to artificial intelligence, quantum computing, and gene editing. Yet, despite these extraordinary achievements, the human mind itself remains largely the same as it was thousands of years ago. This mismatch lies at the heart of many modern struggles.

Unlike technology, biological evolution is painfully slow. Evolution works through genetic changes accumulated over countless generations. The human brain that sends emails and interacts with AI is essentially the same brain that once hunted on the African savannah. Neuroscientists often describe it as a Stone Age brain living in a digital world.

Our emotional circuits evolved primarily for survival, not for handling social media, constant notifications, information overload, or global competition. Fear helped our ancestors avoid predators. Jealousy protected social bonds. Tribal instincts strengthened cooperation within small groups. These traits were adaptive in prehistoric environments, but in today’s hyperconnected world they can fuel anxiety, polarization, hatred, and chronic stress.

Technology, by contrast, evolves cumulatively. Each discovery builds upon previous discoveries, creating exponential growth. Human evolution follows natural selection, while technological evolution follows innovation. One progresses over millennia; the other transforms within decades. This difference creates what some scientists call an evolutionary mismatch—a situation where ancient psychological mechanisms are forced to operate in environments they were never designed for.

Perhaps this explains why scientific progress has not automatically produced emotional maturity. We have developed machines capable of extraordinary calculations, but our brains still struggle with anger, envy, grief, and cognitive biases. The same species that landed probes on Mars can still be manipulated by misinformation, tribal identities, and irrational fears.

Philosophically, this raises an important question: Is humanity suffering from a gap between external evolution and internal evolution? Civilization has accelerated our tools, but not necessarily our wisdom. Knowledge expands rapidly, while self-understanding advances much more slowly.

Neuroscience suggests that reason itself is not the ruler we once imagined. Emotions often guide decisions before logic has a chance to intervene. The rational mind may be powerful, but it is built upon emotional foundations shaped by millions of years of evolution.

Perhaps the next stage of human progress will not come from faster computers or more sophisticated algorithms, but from learning to understand and regulate the ancient emotions that still govern us. Technology may continue to evolve exponentially, but unless our emotional intelligence, empathy, and self-awareness evolve alongside it, humanity may remain caught between the brilliance of its inventions and the limitations of its own nature.

In the end, our greatest unfinished frontier may not be outer space, but the landscape of the human mind itself.

Bringing all five technologies together against the human evolutionary timescale established in section 1: AI’s 300,000-fold compute growth since 2012, at a 3.4-month doubling time, dwarfs not only Moore’s Law’s two-year cycle but every prior technological growth rate in recorded history. CRISPR moved from bacterial-defense discovery to personalized human gene therapy in thirteen years. Google’s Willow chip performed in five minutes what would take a classical supercomputer ten thousand trillion trillion years. None of this has any precedent in human genetic evolution, which moves on a timescale of roughly two million years for a single major adaptive change.

But humanoid robotics and space exploration complicate any simple, uniform story of unstoppable technological acceleration, and that complication is the most intellectually honest finding in this entire article. Tesla’s own admission that its most-hyped robots weren’t doing useful work, and Musk’s own ’50-50′ framing of the Mars timeline, show that even extraordinarily well-resourced technological projects remain genuinely uncertain, genuinely capable of slipping, and genuinely bound by hard physical and engineering constraints in a way that AI’s pure compute scaling, so far, has not been.

The honest synthesis, then, is not “technology has won.” It is more precise and, I’d argue, more useful: technology evolves exponentially within specific, well-defined, often digital or information-based domains, where iteration is cheap and physical constraints are minimal. It evolves far more unevenly — sometimes still remarkably fast, sometimes genuinely bottlenecked — in domains involving physical embodiment, materials, and orbital mechanics. Humans, meanwhile, do not evolve biologically on any timescale remotely comparable to either pattern; human genetic change remains glacially slow by any of these standards. What does move quickly in humans is culture, knowledge, and institutional adaptation — which is precisely the layer increasingly under pressure to keep pace with what humans are building. The deeper question this asymmetry raises is not who is winning a race that was never actually symmetrical. It’s whether the slowest-moving part of this whole system — human governance, ethics, and institutional adaptation — can keep pace with the fastest-moving part: the technology humans themselves continue to build.

What strikes me most, working through these five technologies side by side, is how much the humanoid robotics and space exploration sections actually strengthen the article’s credibility rather than weaken its central claim. It would have been easy to write a piece that simply piles up impressive numbers and lets the reader conclude that everything is accelerating uniformly, forever, without limit. That isn’t what the evidence actually shows.

What it shows instead is more interesting: acceleration is real, but it is domain-specific, and the domains where it is fastest — AI, gene editing, quantum computing — share a common feature. They operate substantially in the realm of information, code, and chemistry, where iteration is cheap, failure is recoverable, and physical constraints are comparatively light. The domains where acceleration is slower and more uneven — humanoid robotics, space travel — are precisely the domains where technology has to contend directly with the physical world’s stubbornness: gravity, friction, orbital mechanics, the genuine difficulty of building a machine that can reliably do what a human hand does without thinking about it.

There’s a connection here, I think, to a much older Indian distinction between the subtle (sukshma) and the gross (sthula) — not as a mystical claim, but as a genuinely useful lens. Information moves at the speed of the subtle: it can double, copy, and iterate with almost no resistance. Matter moves at the speed of the gross: it has mass, inertia, and constraints that no doubling time can simply dissolve. Watching AI’s exponential curve sit next to Starship’s hard-won, setback-punctuated progress is, in a strange way, watching that old distinction play out in real time, in units of months and years rather than millennia.

• Next time you read a claim about a technology’s pace of progress, ask which domain it’s actually in — information/digital (where exponential growth is genuinely common, as with AI) or physical/embodied (where progress is typically slower and more uneven, as with robotics and space travel). The distinction changes how skeptical you should reasonably be.
• If you work in a field being reshaped by AI, take the 3.4-month compute doubling time seriously as a planning horizon — it means capability assessments more than a year old are very likely already outdated, regardless of how recent they felt when you read them.
• Hold Tesla’s own January 2026 admission about Optimus alongside its earlier confident public claims as a working example of how to evaluate any bold technology timeline: look for what the company itself later admits, not just what it initially announced.
• Notice your own reaction to the ’50-50 chance’ framing Musk gave the 2026 Mars timeline. Honest uncertainty stated plainly is a more trustworthy signal than confident certainty — in technology forecasts and in most other domains of life.
• Consider where, in your own work or institution, the ‘human layer’ — ethics review, governance, regulation, simple adaptation time — is the genuine bottleneck, not the technology itself. Per this article’s conclusion, that human layer, not the technology, is usually where the real race is actually being run.

Whose evolution is faster — humans, or the technologies humans build? Having worked through AI’s 300,000-fold compute growth, CRISPR’s thirteen-year arc from discovery to personalized therapy, Google’s Willow chip outperforming a classical supercomputer by a factor with twenty-five zeroes, and the genuinely uneven but still remarkable progress of humanoid robotics and space exploration, the honest answer is that the question compares two processes that were never running on the same track to begin with.

Human biological evolution remains exactly as slow as it has always been — roughly two million years per major genetic change, unmoved by anything happening in a server farm or a cleanroom. What has changed, dramatically and on a timescale with no real historical precedent, is the technology humans design, build, and iterate on, freed from the constraint of generational reproduction that bounds biology. The five technologies examined in this article are not competing with human evolution. They are running an entirely different race, on an entirely different clock — and the genuinely important question, the one this article’s conclusion keeps returning to, is whether the slower, more deliberative layer of human judgment, governance, and institutional adaptation can keep pace with the fast clock humans themselves have built.

1. Sevilla, J. et al. (2022). Compute Trends Across Three Eras of Machine Learning. Foundational paper documenting AI compute growth rate eras, Moore’s-Law-era and Deep-Learning-era doubling times. Visual Capitalist, Charted: The Exponential Growth in AI Computation

2. Medium / Przemek Chojecki (2025). AI Moore’s Law: AI’s Computing Revolution. 300,000x compute growth since 2012, 3.4-month doubling time figures. pchojecki.medium.com

3. Qureshi, N. Moore’s Law For Intelligence. Human Brain Equivalents (HBE) framework and historical/projected HBE growth figures. digitalspirits.substack.com

4. Wikipedia. Technological singularity. Ray Kurzweil’s law of accelerating returns and generalized Moore’s Law framework. en.wikipedia.org

5. NationofChange (2025). Human gene editing and the CRISPR revolution. Doudna/Charpentier 2012 discovery, 2025 CHOP personalized therapy case, US/Europe/China company counts. nationofchange.org

6. Labiotech.eu (2025). CRISPR: The gene editing tool changing the world (2025 update). CRISPR market valuation data, $3.12B 2022 to $4.69B 2024. labiotech.eu

7. Statista. CRISPR genome editing — Statistics & Facts. $33 billion projected 2033 market size; CRISPR mechanism background. statista.com

8. SpinQuanta (2025). Quantum Computing Industry Trends 2025: A Year of Breakthrough Milestones. Google Willow chip specifications, below-threshold error correction, IBM Starling roadmap, IonQ/Ansys result. spinquanta.com

9. Quantum Zeitgeist (2026). Quantum Zeitgeist 2025 Year In Review. Quantum Echoes algorithm 13,000x speedup verification. quantumzeitgeist.com

10. Network World (2025). Top quantum breakthroughs of 2025. Industry-wide 2025 quantum computing milestone summary. networkworld.com

11. New Market Pitch (2026). Figure 03 vs Tesla Optimus Comparison Tracker. Figure AI’s 30,000+ BMW vehicle customer-corroborated data; Tesla teleoperation reporting. newmarketpitch.com

12. BotInfo.ai (2026). Tesla Optimus: Complete Analysis of AI, Specs & Future Outlook. Musk’s January 2026 Q4 2025 earnings call admission on Optimus useful-work status. botinfo.ai

13. Humanoid.press (2026). Humanoid Robots News. June 2026 Figure 03 BotQ production rate; Unitree 2025 shipment figures; Beijing E-Town Half-Marathon “Lightning” robot result. humanoid.press

14. Space.com (2025). Elon Musk says SpaceX will launch its biggest Starship yet by year’s end, but Mars in 2026 is ’50/50′. Musk’s own Mars timeline probability assessment. space.com

15. SpaceOdysseyHub (2026). SpaceX Starship in 2026: Mission Timeline, Tests, and What’s Next. Booster catch milestone, in-orbit propellant transfer test, NASA Artemis HLS 2027 target. spaceodysseyhub.com

16. Aerospace America (2026). A Closer Look at SpaceX’s Mars Plan. Optimus robots’ planned role in initial Mars infrastructure setup; expert skepticism. aerospaceamerica.aiaa.org

17. Rout, N. Yogic Intelligence vs Artificial Intelligence. TheQuestSage.com, Sl 67. Companion piece on the philosophical implications of AI’s growth trajectory referenced in Section 2. thequestsage.com

18. Rout, N. Quantum Computing Explained: 5 Problems. TheQuestSage.com, Sl 69. Companion piece on quantum computing’s remaining open problems, referenced in Section 4. thequestsage.com

P10 — The Next Human: Science, Technology, and the Future We Are Already Building

• Yogic Intelligence vs Artificial Intelligence (TheQuestSage.com, Sl 67) — The book-length companion exploring AI’s outward expansion of intelligence against Yoga’s inward expansion, referenced in Section 2.
• Carbon vs Silicon Intelligence : 5 Fundamental Differences Between Human Intelligence and AI (TheQuestSage.com, Sl 68) — A focused comparison of biological and artificial intelligence at the mechanism level.
• Quantum Computing Explained: 5 Problems (TheQuestSage.com, Sl 69) — The deeper treatment of quantum computing’s remaining open challenges referenced directly in Section 4.
• Generative AI’s Impact on Humanity (TheQuestSage.com, Sl 64) — A companion piece on AI’s societal implications, complementing this article’s technical growth-rate focus.
• The Next Human: 5 Technologies Already Changing Us (TheQuestSage.com, Sl 79) — The broader P10 pillar article this piece extends with a specifically comparative, evolution-speed lens.

---