B13 perspective notes: 2023-12-03 to 2024-02-05 (15 posts)#

These notes read the batch for how Maynard thinks, not for the concepts he names. All fifteen posts were read in full. There are no Modem Futura posts in this batch. Quotes are exact apart from straightened apostrophes, and his typos are kept.

Evidence rules applied

• 2023-12-03 3d-artificial-brains-and-ai reprints his sole-authored 2014 Nature Nanotechnology commentary (“Could we 3D print an artificial mind?”). It counts as his thinking in 2014. His 2023 introduction re-endorses it (“it remains relevant”) and corrects it (“It misses some things”).
• 2023-12-15 pope-francis-artificial-intelligence is mostly long quotation from the Pope’s World Day of Peace message. Only his framing sentences count.
• 2023-12-20 ai-superalignment-and-cooperation-science reports an Unplugged conversation with Mark Daley and Athena Aktipis. Biological claims are theirs (he marks this: “at least as Athena suggested”). His framing and conclusions count.
• 2023-12-22 un-governing-ai-for-humanity quotes the UN report, his own 2008 WEF notes (his), a 2010 proposal written “with my colleague Tim Harper and other Council members” (co-authored) and a 2018 WEF council report (collective). Only the 2008 notes and his commentary count as his thinking.
• 2024-01-17 ai-global-risks-2024-wef-davos quotes the WEF report at length, including two full call-out boxes. Only his framing counts.
• 2024-01-21 how-can-stories-unlock-pathways-to quotes the film The Assignment and Tamara Krinsky. These are not his. The closing passage from Future Rising ch. 29 (“Stories”) is his.
• 2024-01-30 first-in-human-trial-of-neuralink-bci quotes FFTF ch. 7 (Ghost in the Shell) and attaches it as a PDF.
• All header images are AI-generated (Dall-E, Midjourney). There is no AI-written prose in the batch.

Context. The batch runs from a year after ChatGPT’s launch to early 2024: the UN interim report, the WEF Global Risks Report and Davos, the ASU–OpenAI deal, Neuralink’s first implant and the Apple Vision Pro launch. On the surface it is a run of reactions to events. Read for method, it shows him doing the same things again and again:

• going back to his own earlier work;
• testing new technology with his own body;
• building small physical provocations;
• convening unlikely conversations;
• reading institutional risk documents as someone who helped write them.

1. How he thinks here#

He starts from something he did, saw or felt#

The posts open with a personal scene more often than a thesis:

• 2024-02-05 is-the-apple-vision-pro-a-social-game-changer: “11 AM Sunday morning and I’m queuing outside the Scottsdale Fashion Mall Apple Store. It may be the nerdiest thing I’ve done in a while!”
• 2024-01-07 the-future-of-being-human-is-analog: “Walk down Cady Mall on Arizona State University’s Tempe campus and, if you’re observant, you may catch a new addition”.
• 2023-12-20 ai-superalignment-and-cooperation-science: “It was one of those double-take moments in a conversation where two seemingly different worlds meshed”.
• 2024-01-21 how-can-stories-unlock-pathways-to: “I must confess — I’m a sucker for a good dystopian sci-fi movie.”
• 2023-12-03 3d-artificial-brains-and-ai: “Nine years ago I published a highly speculative commentary”.

The analysis comes after the encounter. The Vision Pro post moves from a queue, to a history of sci-fi expectations, to a technical reframing, to first-hand tests, to published research, to privacy, to what might happen to society.

He finds out by using the thing, playfully and on purpose#

The Vision Pro post is the clearest example in the batch of play used as a way of finding out.

• He calls his motive curiosity about how the device “might change how users interact with virtual and real environments — and how this in turn might change them”.
• He sets himself tests. He reads a paper book outdoors through video passthrough as “a good test of the resolution of the system”. The book is Ready Player One, and he watches the film inside the headset at the same time, a joke for “readers who know something of the history of VR in science fiction”.
• He stacks one experiment on another. He takes a driverless Waymo while wearing the headset: “I thought why not combine two cutting edge technologies”. He reports the result as an experience, not a spec: “there’s something quite deliciously uncanny about placing a large movie screen right in front of your self-driving car.”
• He treats himself as a data point: “despite being susceptible to motion sickness, I had no problems”.
• He then checks his experience against a Stanford paper (“I can certainly attest to that”) and criticises the paper’s method. Its conclusions are probably driven by the Quest 3, so “research using the Vision Pro may throw up different results.”

The order is: play, then observe, then compare with evidence, then critique the evidence. Delight and rigour are part of the same inquiry.

He builds things so that others will think#

The Vestaboard post (2024-01-07) is about an installation he made. A split-flap board cycles through “over 800 quotes, questions and thoughts associated with the future of being human”, one every ten minutes. The design choice is itself the argument. He wanted “a technology that highlighted rather than subsumed our humanity”. It is “a ‘slow technology’ — it’s made to be savored”. And: “This is a technology that makes you think, rather than doing the cognitive heavy lifting for you.” Written a year into ChatGPT, that line is a quiet counterpoint to AI. It is made through an object, not a polemic.

He also frames the installation as an open experiment whose result he does not know: “I’m intrigued to see if others feel the same … Or maybe not.”

Plausibility tests on speculative technology#

The 2014 commentary (reprinted 2023-12-03) shows the physicist’s move that runs through his work:

1. Take a far-out possibility, a 3D-printed “artificial mind”.
2. Ask first “whether there is a plausible likelihood of such a convergence occurring in the first place”.
3. Work through the engineering by structural analogy with the brain. Heat removal is the hard problem, and “In the human brain this is achieved predominantly through blood perfusion”, so printed microfluidic channels might do the same job.
4. Name the limits: “Of course, this is a naive thought experiment, and one that glosses over many substantial technological challenges.”
5. Turn at the end to responsibility: “if our technological capabilities are beginning to shift from the fanciful to the plausible in constructing an artificial mind that has some degree of awareness, how do we begin to think about responsibility in the face of such audacity?”

The Vision Pro post uses the same test against science fiction. The seamless blending seen on screen would need an overlay projected into the eye or “a virtual feed beamed directly into our brains — something that … we’re not going to see in my lifetime.” Sci-fi sets the expectation (“we’ve been conditioned to take it for granted”) and physics checks it.

Reframing by inverting the structure#

• Vision Pro: “Rather than overlaying a virtual world over real life like the HoloLens, the Vision Pro overlays real life over a virtual world”. Once reality becomes a digital reconstruction, it “can now be manipulated in ways that physical reality simply cannot”. This inversion is the insight the rest of the post hangs on.
• Superalignment: a problem framed as controlling a smarter machine becomes a question about “hierarchical multi-agent systems” in which “‘less intelligent’ agents manipulate the behavior of ‘more intelligent’ agents to their benefit”. The frame is set “irrespective of the ‘substrates’ within which it resides.”
• WEF 2024 (2024-01-17): a report the headlines read as an AI alarm is reframed by its own rankings: “If anything, this report is a call to action on climate, not AI.”
• Modesta (2024-01-28): in zero gravity “everyone starts from a point of ‘disability’”, so ideas of an “able” astronaut “are merely constructs of our own biases.”

“What if” and “imagine”#

He often reaches for a scenario to open a question:

• “Imagine, for instance, workspaces — or coffee shops — that are located on opposite sides of the world, yet have identical physical layouts” (Vision Pro).
• “(Imagine, for instance, a future where a BCI is a prerequisite for some jobs … or some educational programs)” (Neuralink).
• The whole 2014 piece is a “what if”.

Serendipity sought and prized#

The superalignment post puts it plainly: “one of those serendipitous moments that academics like me live for, where ideas from different disciplines suddenly align and new insights begin to emerge.” Unplugged is designed to produce such moments: “to tease out new insights and ideas through casual and candid conversation”. He enjoys where the talk wandered (“we ended up — bizarrely but appropriately — on what AI developers can lean from kombucha”). He plays with an analogy while marking its limits: “humans acting as metaphorical viruses in a world of superintelligent AIs … although I’m not sure how far anyone would want to push this analogy.”

Quick, crude, public analysis driven by curiosity#

In 2024-01-14 wef-global-technology-risk-trends he tabulates 18 years of WEF technology risks because “I wanted to simply see”. He is frank about the method (“This is an exceptionally crude way of approaching the data, but one that I think is still insightful”) and says a proper study “would be worth someone spending some time doing at some point.” He later posts a correction when his table had errors. Curiosity sets the pace, and honesty about method and error keeps it credible.

Holding outcomes open, and ending on a question#

He often declines to close:

• “How this will play out is anyone’s guess at this point. It may be that evolution will trump technology at every turn … But it’s equally possible” (2024-01-01).
• Organoids: “what happens if these bio-AI’s develop consciousness. And if they did, how we’d know …” (2023-12-13).
• Neuralink: “we still have very little idea what the risk landscape will look like around devices that allow your brain to be wirelessly connected to the internet via your phone …”
• Vision Pro: a society “living in a reality that is no longer constrained by … well … reality.”
• WEF: “As they say, watch this space …”

The trailing ellipsis is a habit. It leaves the reader holding an open question, not a verdict.

2. What matters to him#

What it means to be human, including what we might lose without noticing#

The New Year post (2024-01-01) is a statement of purpose. He is “fascinated by how we define ourselves by the technologies we develop” and “intrigued by how emerging technologies enhance or diminish what we consider to be the essence of what it means to be human.” Then: “what drives my work more than anything is the possibility that we’re reaching a tipping point where, rather than being a means of augmenting and expressing who we are, our technologies begin to fundamentally change who we are — or even what we are.”

In the Pope post he says what he is protecting. The Pope’s lines on mortality and “technological dictatorship” matter “when approached from the question of what makes us human — what defines us — and understanding what’s worth protecting lest we discard it without realizing what we’re losing, these ideas resonate deeply with me and my own work.” The danger is not only harm but losing something valuable without realising it.

Agency and consent#

Emerging technologies may lead “to alterations in this base code without our agreement or permission” (2024-01-01). The same concern shapes the Neuralink post. Musk’s fans imagine superhumans “locked into a subscription plan that governs what they can do and what they cannot.” Autonomy is the value at stake in both.

Dignity, disability and who counts as “able”#

• The Mission: Interplanetary disability episodes “changed my perspective of human in space more than anything else we did” (2024-01-28).
• He admires Modesta as someone “willing to throw away the constraints of conventional thinking.”
• He publicly corrects “suffering” from quadriplegia after Gregor Wolbring pointed out it implies those “not considered ‘human-typical’ are suffering and need to be fixed — which is clearly not right or appropriate” (2024-01-30).

Inclusion and the public interest#

Among his UN takeaways (2023-12-22):

• concern about the “AI divide” and the Global South;
• AI for social good “rather than simply for financial gain”;
• “I’m an ardent advocate for multi stakeholder engagement around advanced technology transitions”;
• “AI for good cannot simply be left in the hands of tech entrepreneurs and AI experts”.

Curiosity, and the fear of losing it#

The Vestaboard post’s most revealing line is a worry: “It’s entirely possible that there’s simply not enough curiosity left in the world for people to wonder what the clackity machine does”. Curiosity is something he both practises and wants to protect.

Delight#

The batch is full of pleasure, which is itself evidence of how he works:

• “And I’m just a little in love with it!” and “I am a little obsessed with the sound of the board” (Vestaboard);
• “at least I have the analog joy of being surprised by what comes up”;
• “If you thought that 2023 was a far-out year for AI, it’s about to get weirder!” (2023-12-13);
• “I was blown away — unexpectedly so” (Vision Pro);
• “riveting from beginning to end” (Unplugged);
• “I’m very excited to see where it leads!” (ASU–OpenAI).

What frustrates him#

• People who “speak with authority on what AI might do without thinking about what it actually is” (Pope post).
• The belief that good intentions will produce beneficial AI: “this is sheer fantasy — but it’s a fantasy that a surprising number of people seem to have bought in to” (Pope post).
• The assumption “amongst some AI developers that the market around transformative and disruptive AI will automatically deliver on social good” (UN).
• Risk agendas “driven more by headlines than foresight” (2024-01-14).
• Crowdsourced conclusions where “innovative and creative ideas are likely to be subsumed by regression to the mean of conventional thinking” (2024-01-17).
• Flashy technology: “the older I get the less interested I find myself in tech that’s flashy for the sake of being flashy”.
• Small practical failures matter too. The podcast episodes appeared in reverse order, “rather irritatingly”, so “I fixed things!” (2023-12-06).

Education as the long, hard work#

Teaching students to “grasp the social and ethical aspects” of technology is “so important — and yet so hard to achieve”. SFIS “was set up for precisely this purpose”, and “It’s been an uphill battle” (Pope post).

3. Risk as a way of thinking#

The terms “risk innovation” and “orphan risk” do not appear in this batch. “Risk landscape” appears twice and “navigate” appears in his own voice. The deeper pattern is how he reframes what counts as risk.

“Risk landscape” and “navigate” in use#

• “we still have very little idea what the risk landscape will look like around devices that allow your brain to be wirelessly connected to the internet” (2024-01-30). The landscape is unmapped territory, not a list of hazards.
• Decision-makers “would be ill-prepared to understand and respond to an evolving global risk landscape” without the WEF report (2024-01-14).
• “we need to be thinking hard and deep about how this tech might impact our lives, and how to successfully navigate the technology transition Apple is ushering in” (2024-02-05).
• “we’re entering new territory with the technology” (2024-01-18).

A small marker of change: his own 2008 WEF notes call for policies “to foster and manage the effective and sustainable development of emerging technologies” (quoted in 2023-12-22). In 2024 his own verb is “navigate”. This is one instance, but it fits the shift he describes.

Threat to value#

The Pope post’s “what’s worth protecting lest we discard it without realizing what we’re losing” is risk as a threat to value in plain words. What is threatened is not health or safety but what defines us. Even his New Year greeting is phrased in terms of value: “here’s to a 2024 where the present of being human is one that is full of value to you — however that comes — and one sets the stage for an even more value-filled future of being human.”

The risks he raises in this batch are nearly all threats to something people value, and none fits hazard-and-exposure assessment:

• identity and the “base code” changed “without our agreement or permission” (2024-01-01);
• BCIs setting new norms for “who gets to be be ‘augmented’ and how this impacts everything from education to jobs” (2024-01-30);
• augmented humans on subscription;
• the need for “some recalibration of how we collectively think about privacy” once people wear sensor-laden headsets (2024-02-05);
• how immersive technology “potentially messes us up if we’re not prepared” (2024-02-05);
• the moral status of organoid AIs “and if they did, how we’d know” (2023-12-13);
• responsibility “in the face of such audacity” (2014 commentary).

A new kind of technology needs a new way of thinking#

The New Year post makes the case without calling it risk. Most past technologies were “extrinsic to our sense of self”; few were “intrinsic”, changing us “from our constituent molecules and cells all the way up to our consciousness”. Emerging technologies may let us alter “the intrinsic ‘base code’”, and AI’s “increasing mastery over language” is part of that. The result could transform being human “in ways we haven’t previously experienced as a species.” His response is not a new procedure but “a greater need than ever for new thinking” — and he admits “we still have a long way to go in terms of even understanding the questions we should be asking here, never mind finding answers.”

He also stresses that the technologies “defy convenient categorization” because they converge (nanoscale science, quantum, gene editing, synthetic biology, neuromorphic computing, BCIs, “and even social media”). Things that cannot be categorised cannot be handled by category-based risk tools.

Over-the-horizon risk and the limits of expert crowds#

The two WEF posts show him reading an institution’s risk-sensing apparatus as a risk scientist:

• The report “tends to reflect the current ‘risk perception zeitgeist’ rather than over the horizon emergent risks.”
• He worries about “over the horizon risks that, if not mapped out, are the ones most likely to take us unawares.” This is the logic of orphan risks without the name.
• He wants “more foresight methodologies to augment the wisdom of the expert crowd”. Foresight is added to expertise, not substituted for it.
• He reads rankings closely (AI 29th of 34 short-term, 6th long-term) and spots artefacts of categorisation. The jump in AI-related concern was “aided perhaps by risks associated with ‘misinformation and disinformation’ … being moved from the area of societal risk and into technological risk”. That is a trained quantitative risk analyst at work, building on his foundations rather than discarding them.
• He values the report’s move to “approaching sociotechnical systems from a risk perspective, rather than simply focusing on technologies alone.”

Risk of not innovating well, and too much risk aversion#

• At the Project on Emerging Nanotechnologies, “we were front and center of global discussions around the potential dangers of not developing nanotech responsibly” (2024-01-14). The risk is framed around getting development right, not around the material alone. That foreshadows risk innovation.
• From Unplugged he takes away that “threats to cooperative systems build resilience and immunity — a concept that suggests too much risk averseness is a bad thing” (2023-12-20).

Mental models that open possibilities#

In this batch the load-bearing ideas are ways of seeing:

• extrinsic versus intrinsic technologies;
• the “base code”;
• AI as a mirror, “albeit a rather imperfect one”;
• alignment as cooperation in a hierarchy, whatever the substrate;
• “slow technology”.

None is an operating procedure. Where he does talk method (WEF), he asks for a broader toolkit, not a new checklist. In the UN post he credits the report for being “neither a report that gets tangled up in speculative long term AI risks, nor misses the transformative significance of AI breakthroughs”. That balance keeps both risk and possibility in view.

Humility#

• “(although I may well be wrong here)” (Vision Pro).
• “it’ll be a close call whether the Spatial Computing revolution is upon us now, or whether Apple have simply lit a long but hard to extinguish fuse.”
• “many of the fears and aspirations that AI has sparked are more speculation than imminent reality” (2024-01-01).
• He admits his own past blind spots. Pervasive computing was “probably not something I would have pegged as a global risk at the time”, and nanotech’s long run on the list came “at the expense of other disruptive emerging technologies”, a field he himself was part of (2024-01-14).

4. Scholarship and public writing#

His own account of his method#

The Moviegoer’s Guide post (2023-12-06) says what Films from the Future was for: “The big thing I set out to do with Films from the Future was to weave together a tapestry of ideas that draw on a myriad different disciplines to create a unique perspective on technology innovation and the future.” He calls the podcast “a rather idiosyncratic (and hopefully engaging) journey” and “a surprisingly deep dive”. Idiosyncrasy, engagement and depth are named together as aims.

He also says his writing has a voice meant to be heard: “I write with a very specific internal voice”, and the podcast lets you “hear the book as it exists in my head.” Accessibility is a practical aim: “the easiest way to access the book Films from the Future for free!” and “The podcast’s made for ears.”

Continuity: public writing as a living archive of his scholarship#

• He reprints his 2014 Nature Nanotechnology piece in full, noting that “At the time it was so edgy that my editor at Nature Nanotechnology was unsure whether the commentary was appropriate”, and that it now looks “increasingly prescient”. He also says what it got wrong: “It misses some things — the need for a secondary network to systematically adjust weights between neurons”.
• He places the UN report in a lineage going back to his 2008 WEF notes and the 2010 Davos proposal, linking the original PDFs. He is careful about credit: “There’s no concrete evidence of causal links … apart from my continued input.”
• For Neuralink he goes back to FFTF ch. 7 and attaches it.
• He closes the stories post with Future Rising ch. 29: stories “become the pivot point between being able to imagine the future and beginning to build it” and “help form the mental maps that chart a way between past, present, and future.”

Film and story as teaching tools#

He writes to readers the way he teaches: “As I’d tell my students if we were watching this together, it’s worth paying attention to the cinematography, the ‘mise-en-scènes’” (2024-01-21). He loves dystopias because “they force us to think about what it means to be human when everything around us is falling apart”, yet he was “surprised by just how much I was captivated” by a non-dystopian film. In the Vision Pro post, sci-fi (Ready Player One, Minority Report, Black Mirror) is the frame for public expectation, and he points to written sci-fi (Toby Weston’s Singularity’s Children) as exploring more transformative scenarios than film does.

Evidence and expertise#

• He goes to primary sources and sends readers to them. For the Vatican text: “I’d encourage you to read the message in full.” He also gives links to the UN report, the WEF report, the Nature Electronics paper and the Stanford paper.
• He separates source from coverage. Headlines suggest catastrophe, “And yet the focus on artificial intelligence in the risk report is more nuanced and less sensationalist” (2024-01-17).
• He criticises method on both sides: the WEF’s crowdsourcing and the Stanford paper’s reliance on the Quest 3.
• He praises critics in their own terms. The Stanford authors suggest more research “as any good researcher should”.
• He corrects in public (the WEF table, “suffering”).

Transdisciplinarity in practice#

• Unplugged puts a cooperation biologist and a Chief AI Officer in the same room.
• He welcomes the Pope’s call for “cross-disciplinary dialogue (something I have been calling for for some time now)”.
• He works with screenwriters (Scirens), a bionic pop artist, space-accessibility advocates and a former astronaut.
• He notes “all four authors are women, and none are from North America or Europe — reflecting the importance of listening to diverse voices tell stories of different futures.”

Craft#

• He puts a tl;dr up front (Vision Pro, Modesta).
• He explains ideas in plain terms: the von Neumann bottleneck, optical versus video passthrough.
• He uses light humour (“pun intended”; “the platform formerly known as Twitter”).
• He knows about narrative shape. The stories post has an Afterword for material “which would have diluted and disrupted the piece’s narrative if I had — so I dumped it all here”.

5. His role as he sees it#

An insider who stays a critical friend#

He has been inside the institutions he writes about:

• chief science advisor to the Project on Emerging Nanotechnologies;
• chair of the WEF Global Agenda Council on Emerging Technologies (2011);
• a member of the Global Future Council on Agile Governance;
• “regularly asked to contribute” to the Global Risks Report.

He still calls that report “in some ways, a flawed report”, its recommendations “rather lackluster”, and its AI literacy proposal something that “barely scratches the tip of the iceberg”. He is fair to it too. In an addendum he writes: “having accused the Global Risk Report of ‘regression to the mean of conventional thinking,’ I though that the two ‘next global shock’ call-outs … were actually rather insightful.”

Convener and builder of spaces#

He hosts Unplugged, invites readers to the Modesta conversation (“hope you’ll join is”), installs the Vestaboard, and runs the Future of Being Human initiative. The New Year post presents the initiative as a response to a question, not as a centre with answers: “we still have a long way to go in terms of even understanding the questions we should be asking here”. He makes places where others can think, and does not only deliver conclusions.

Readers as fellow explorers#

• He recommends things while disagreeing in part: “You may not agree with all of it (I don’t)” (Pope post).
• He speaks to readers across belief: “Irrespective of whether you are religious or not”.
• He is modest about his own work: “It’s always a little uncomfortable suggesting people listen to your own work”.
• He is open about his biases: “this may just be my age and inclinations speaking … But I suspect there’s something to human-scale technologies like this that transcends my own biases”, and “I’m still a IRL kind of person”.

How he treats those he disagrees with#

• Musk. He criticises the “cult of personality” and the old hiring line “No neuroscience experience is required”. He also credits the medical promise (“it could be groundbreaking”) and stays “intrigued to see how this technology evolves.” There is no caricature.
• The Pope. He engages with a religious text on its merits and marks his own difference: “I tend to go beyond the limitations of reproducing human abilities in how I describe AI”.
• Critics of the Pope’s view. “I suspect some will disagree that this is, indeed, a risk.”

What he refuses to do#

• Preach. The stories post states his theory of persuasion: “Preach to someone about the future, and most people will shut down. Beat them over the head about how their ideas and thinking about the future are wrong, and they’ll likely fight back — or simply ignore you. But tell them a story that resonates with them, and you open their mind to exploring possibilities that they would otherwise reject out of hand.”
• Sensationalise. He rejects the Davos “global AI catastrophe” headlines.
• Hype. On Neuralink: “we’re decades — generations possibly — away from brain-computer interfaces that live up to the aspirations of Musk’s followers.”
• Give in to techno-solutionism. Good intentions alone are “sheer fantasy”.

Enthusiast for hands-on access#

The ASU–OpenAI post (2024-01-18) is his least hedged. He was among those “clamoring for access … myself included”, expects “a tsunami of creativity and innovation”, and relies on “the right checks and balances” with both parties “being diligent here”. This is institutional news from inside his university, and his critical apparatus is light. It does show a consistent disposition: he wants creative people to be able to use a technology in order to learn what it is for, and puts access-to-experiment first.

Changes of mind and surprise, shown openly#

• The disability episodes “changed my perspective”.
• The Vision Pro: “blown away — unexpectedly so”.
• The Assignment: “surprised by just how much I was captivated”.
• He corrects his own 2014 piece, his WEF table and his Neuralink wording.

6. What is distinctive#

1. Thinking about AI through its physical substrate. A physicist and nanoscientist, he asks whether the next step is “a revolution in the substrates these run on” (3D neuromorphic printing, organoid “Brainoware”). He links the physics straight to consciousness and responsibility. Few AI commentators join materials, compute architecture and moral status in one line of thought, and he did it in 2014.

2. Alignment through the science of cooperation. He takes superalignment into evolutionary biology. In robust hierarchies, less intelligent agents steer more intelligent ones through aligned interests and information flow, and intelligence may be “not necessarily the primary purpose or the pinnacle achievement of the system”. “Human exceptionalism” is named as a blind spot. This substrate-independent, cooperative framing differs from both control-based AI safety and dismissive scepticism.

3. AI not treated as a copy of us. Several posts point the same way. The 2014 piece imagines brains “inspired by our own, but do not necessarily emulate them”. He goes “beyond the limitations of reproducing human abilities” in defining AI. Machines emulate what makes us unique “in our own eyes”. This leaves room for AI to be a new kind of thing, which is why old categories may not fit.

4. Being human as the unit of concern. The extrinsic/intrinsic distinction, and language as part of the “base code” of self and social identity, put AI’s deepest risk in identity and formation, not in catastrophe or narrow harm.

5. His own body and playful experiment as instruments. Reading Ready Player One inside a headset while watching the film, and riding a driverless car in blended reality, then checking the experience against published research. Play produces the observations.

6. Designed objects as arguments. A 132-character mechanical board, “a technology that makes you think, rather than doing the cognitive heavy lifting for you”, is a public intervention in the AI moment made through slowness and delight rather than critique.

7. Reading the risk-sensing system from inside. His 18-year WEF analysis diagnoses how institutions notice risk (zeitgeist, headlines, regression to the mean). He was part of the nanotech wave the reports recorded, and he admits his own blind spots.

8. Disability as a lens on augmentation. Zero gravity as the place where “everyone starts from a point of ‘disability’” turns the BCI and enhancement debate from fixing deficits to questioning norms.

9. Stories over sermons, dystopia and hope together. He loves dystopias as thinking tools and argues that stories open minds where preaching closes them. He does not choose between the two.

7. The posts in this batch that best show how he thinks#

1. 2024-02-05 is-the-apple-vision-pro-a-social-game-changer. Play and first-hand experiment as inquiry. Sci-fi as the frame for expectations, checked by physics. A structural inversion (reality overlaid on the virtual). Critical reading of research, open uncertainty, and “navigate the technology transition”.
2. 2024-01-01 the-future-of-being-human-in-2024. His statement of purpose (“what drives my work more than anything”). The extrinsic/intrinsic model, language as “base code”, consent, open outcomes, questions before answers, and value as the closing note.
3. 2024-01-07 the-future-of-being-human-is-analog. Delight, curiosity and building. A slow, human-scale technology designed as a provocation. Self-aware about bias, and treated as an experiment with an unknown result.
4. 2023-12-20 ai-superalignment-and-cooperation-science. Serendipity he says academics “live for”. Transdisciplinary conversation as method, analogy by structure with its limits flagged, a substrate-independent reframing of alignment, and doubt about excessive risk aversion.
5. 2024-01-14 wef-global-technology-risk-trends. The risk scientist reading how institutions sense risk. Crude analysis done in public and corrected. Insider-critic stance, and over-the-horizon risk versus the zeitgeist.
6. 2023-12-03 3d-artificial-brains-and-ai. The 2014 plausibility thought experiment. Physicist’s substrate thinking, analogy with brain physiology, honest limits, a turn to “responsibility in the face of such audacity”, and a 2023 self-correction.

Close runners-up: 2023-12-15 pope-francis-artificial-intelligence (“what’s worth protecting lest we discard it without realizing what we’re losing”; generous engagement across difference) and 2024-01-21 how-can-stories-unlock-pathways-to (his theory of persuasion: stories, not sermons).