JANUARY - 2026
[on Year 2025]
Farewell to 2025. One book and 2,001 covers. Much to be done.
[on a SF EA Global]
"No trumpets sound when the important decisions of our life are made. Destiny is made known silently"
(Agnes De Mille)
I attended a momentous EA Global in San Francisco. AI safety folk were much in evidence. I did my spiel on the binding problem. Digital mind is an oxymoron. LLMs and their agential offspring are
insentient ignoramuses that impressively outperform pre-neurochipped humans in many cognitive domains but won't be staging a zombie coup. Yet for me at least, the defining moment of the conference was being offered a 300mg gabapentin (Neurontin) capsule. Thank you Jordan. I'm still not entirely clear why I decided to take it. I don't try random pills. Gabapentin had struck me as too uninteresting to be worth even noting in the Good Drug Guide. Despite its suggestive name and structural similarity to GABA, gabapentin doesn't bind to GABA receptors. Nor does gabapentin directly affect GABA uptake or synthesis. Instead, gabapentin binds specifically to the α2δ-1 subunit of voltage-sensitive calcium channels, decreasing their density on the neuronal cell surface. Or at least, that's how the usual neurobabble goes. By blocking the α2δ-1 subunit of voltage-sensitive calcium channels, gabapentin reduces the release of excitatory neurotransmitters implicated in pain signalling, notably glutamate and substance P. The pain of my own NU existence has been almost entirely psychological. My drug regimen has been constant for over two decades. Since autumn 1995, my ancestral namesakes have all taken selegiline. As Wikipedia notes, I wrote the HI manifesto in six weeks after starting selegiline. Since 2002, I've also taken c.200 mg amineptine daily. Optimising my diet, sleep and exercise routine like a naturopath has doubtless helped too. But I've now added 12 x 300mg gabapentin to my regimen. Gabapentin has let me break my post-Covid topical steroid addiction (TSWS). In addition, using gabapentin takes the anxiogenic edge off my predominantly dopaminergic regimen. Gabapentin also makes me feel somewhat "euthymic", as psychiatrists say, an everyday psychosis mediating (what passes for) mental health in several billion natalist human primates daily worldwide. Of course, a measure of tolerance to gabapentin's effects soon sets in. Back in Portugal, I'm a fairly troubled soul once more, brooding on suffering and Everett. But tolerance is incomplete. I feel more mellow. My regimen has changed.
But there was more to SF than my gabapentin epiphany.
I gave a talk on avoiding ethical catastrophes (pdf) at Sentient Futures. Previously, I'd explored how Rare Earthism - my own tentative view - could be wrong. If so, then post-suffering Earth-originating life has an obligation to launch cosmic rescue missions to fix other Darwinian hellworlds within our cosmological horizon. Here at Sentient Futures again, I tried (more lamely this time) to steelman what I don't believe, namely that AI run on classical digital computers - and indeed futuristic non-biological quantum computers - can support phenomenally bound minds with a pleasure-pain axis. Your notional digital "whole-brain emulation" is a zombie. I'll be back in the Bay Area in late May for the inaugural Machine Consciousness conference at Berkeley if anyone would like to plot the future of sentience in our forward lightcone; or hang out.
[on gabapentin]
Describe the subjectively enjoyable effects of high-dosage gabapentin reported by some users:
Taking gabapentin can induce feelings of mild euphoria in a minority of subjects. Discuss possible mechanisms:
[on ending suffering]
On stage with Anders Sandberg and Sarah Wilson at the Institute of Art and Ideas event:
X : mp4
Would Buddha have pressed the OFF button?
The trouble with invoking Buddha is one can be sure that someone in the audience will most likely say you haven't understood the True Meaning Of Buddhism. Maybe.
And here is a YouTube podcast with Micah Zarin.
"Should We Abolish Suffering?"
Preview:
YouTube :
mp4
Full:
YouTube :
mp4
Highlights:
DP and Micah
"There are no experts on consciousness..." (mp4)
DP and Micah
"'We live in Hilbert space'? Ah. Yes, I’ve never known anything else..."
(mp4)
DP and Micah
"We can engineer children with extremely high hedonic set-points..." (mp4)
DP and Micah
"Humans are the worst Darwinian malware. And yet..." (mp4)
DP and Micah
"Don't explore psychedelia until we've sorted out our reward circuitry..." (mp4)
DP and Micah
"Ending predation - the idea that sentient beings shouldn't harm each other will one day seem axiomatic..." (mp4)
DP and Micah
"We need to devise a new signalling system - life based on gradients of bliss..." (mp4)
Imagine if new computer code were conceived in the heat of sexual passion.
Crazy.
Alas genetic code for sentient beings gets chosen in the height of sexual ecstasy. I hope tomorrow the code for all sentient life can be chastely designed to the same exacting specifications as Tesla cars.
Robert, I think in one sense suffering can be extensively studied only by not deeply, empathetically understanding what one is investigating in any but a shallow formal sense. If one really understood suicidal despair or cluster headaches in any but a detached abstract way then one would go insane. Even now, I don't really understand what I'm nominally writing about. So exactly how does one optimise the mix of selective ignorance / knowledge to be maximally productive?
I agree with you about the need for an algonomy - though algonomy is a science that (uniquely) can I hope one day be forgotten.
One query though. Does the problem of suffering really fascinate most people? From my experience, most people try not to think about suffering. It's a turnoff. One of the reasons that I focus so much on the vision of a glorious blissful future is that naïve suffering-focused ethics can be depressing. One loses most of one's audience.- and an exponentially more brutally competitive battle for attention is unfolding.
So what's the right balance for the abolitionist project?
I don't know.
And kudos to Andres and his team for spinning off ClusterFree from QRI.
A Hundred Year Plan to Fix the Problem of Suffering: The 100 Year Plan.
For expository convenience, some details are omitted.
Extreme hyperthymics? A few have come forward. I hope their close biological relatives can have their full genome sequenced too. All prospective parents should be offered the opportunity to choose low-suffering genomes for their future children.
Such a program would be extremely cost-effective.
I suspect more extreme hyperthymics still await discovery, unknown because - like Jo Cameron - they think themselves "normal".
In the immense state-space of qualia, most of which are intrinsically neither hedonically good nor hedonically bad, evolution via natural selection stumbled on the pleasure-pain axis and "encephalised" it. The contents of our phenomenal world-simulations (aka "perception") accordingly feel good or bad in (typically) fitness-enhancing ways - a bland description that fails to evoke the indescribable horrors of severe suffering.
However, a species has evolved, Homo sapiens, that can imagine more civilized signalling systems. I've focused on blueprints for engineering a pleasure-superpleasure axis - a rich information signalling system driven by gradients of intelligent bliss, with a nod to smart neuroprostheses.
But do you think there are others?
Kind? Bless you. Yes, I try to be kind. But fixing the problem of suffering will call for a mix of remorseless hyper-systematization and Super-Machiavellian political guile.
[on ecstatic epileptic seizures]
The terminally ill, the neuropathic pain-ridden, the suicidally depressed - but really, all sentient beings - deserve to taste the divine:
Ecstatic Seizures
("Some People Experience Blissful Ecstasy Right Before a Seizure. Could Understanding This Feeling Help Treat Depression?")
Understanding epileptic seizures promises to unlock the molecular signature of pure bliss - a momentous discovery in the history of science, ethics and medicine. This phenomenon shouldn't be treated as niche, but mainstreamed.
And why optimize only hyperactivation of the dorsal anterior insula?
Superintelligence should optimize all accessible matter and energy likewise.
[on fun]
Wild mice, rats, shrews, and even frogs and slugs can enjoy life to the full :
Jungle Gym
("Wild Animals Run on Wheels for Fun")
It would be easy to treat this result just as an amusing curiosity.
But imagine replacing “Nature, red in tooth and claw” with a biosphere based on fun.
Impossible?
Not with biotech and ubiquitous AI harnessed to engineer a pan-species welfare state.
All sentient beings deserve fun.
[on abolitionist AI]
I was completely blindsided by the transformer revolution in AI. A tantalising prospect now arises. Could abolitionist AI fix the problem of suffering this century rather than over hundreds of years as I anticipated in HI? Could a signalling system based on genetically programmed gradients of bliss be engineered over decades not centuries?
Alas, the socio-political obstacles make me fear such a vision is utopian. But if a breakthrough into the socio-political mainstream occurs (how?), the technical obstacles to heaven on Earth are far less daunting than they seemed 30 years ago.
Modifying a dozen or so genes (SCN9, FAAH, FAAH-OUT, SLC6A4, BDNF, FKBP5,
OPRM1, COMT, CRHR1, HTR2A, OXTR IL6/TNF) could defang and even trivialise unpleasant experience. Archaic biological robots can glimpse if not grok the prospect.
Can AI do the equivalent of AlphaFold for designing alleles / allelic combinations mediating lifelong bliss - or better, elevated hedonic set-points that deliver lifelong information-sensitive gradients of bliss?
I believe so. But how can a biohappiness revolution go mainstream?
See too:
On the Future of Species - unnatural selection
("Geneticist and entrepreneur Adrian Woolfson argues that genome engineering and AI will let us design organisms beyond nature’s limits")
Compare C. elegans. This little millimeter-long nematode is a conscious being like you or me. C. elegans is a phenomenally bound subject of experience who loves morphine. Sure, with only 302 neurons, C. elegans won’t be exploring higher category theory. But C. elegans has a mind, whereas superintelligent AI run on a classical digital computer is a zombie. Binding is the computational-functional superpower of animal minds. It’s non-classical. Binding grants access to the empirical realm of conscious mind forever impenetrable to digital zombies.
Morphine-loving C. elegans
("Caenorhabditis Elegans Exhibits Morphine Addiction-like Behavior via the Opioid-like Receptor NPR-17")
[on thinking]
Shivon Zilis wtites:
"When you think, what medium do you tend to think in?
Would be very curious to hear how you’d describe the base unit(s) of your thoughts and how they feel to you.
I assumed what happens in my head was similar to everyone else but have been surprised by how varied thought can be"
The phenomenology of my thought-episodes is thin, subtle and elusive. I lack the vocabulary within our everyday conceptual scheme to describe it. Disconcertingly, this subtle phenomenology changes if I take a psychedelic - again in ways that transcend my ability to describe. But such generic changes hint at how the nameless phenomenal medium or vehicle of our thoughts shapes their ostensible content - and hence one’s entire conception of reality.
[on the Singularity]
"We have entered the Singularity"
(Elon Musk)
with just as much pain and suffering as before - rampant tribalism, squabbling nation-states, and the horrors of factory-farming expanding?
Same old core emotions, same hedonic treadmill, same hedonic range (etc) ...I'm as bowled over by the transformer revolution in AI as anyone. It's astonishing. But until subjective quality of life is improved for most if not all sentient beings, then it's all just rearranging the deckchairs.
Roll on a true Singularity - i.e. the transition to a transhumanist “triple S” civilization of superintelligence, superlongevity and superhappiness.
Full-spectrum superintelligence entails superhuman mentalising prowess - i.e. a rich empathetic understanding of the perspectives of all sentient beings.
It’s easy to be dismissive and reject such prowess as “woke”.
A high load of "autistic" Neanderthal DNA and a Y chromosome will flatter anyone who takes an "IQ" test; but such tests reveal more about the intellectual limitations and cognitive style of the Caucasian male Aspergers who design them than they do about general intelligence.
On what comes next?
Even superintelligent digital zombies can’t overcome their hardwired ignorance of what they lack, i.e. phenomenally bound minds and access the (mostly still unexplored) empirical realm. I anticipate that advances in neurochipping technologies will shortly allow sentient (trans)humans to embed zombie superintelligence in their CNS. Genetically rewritten, hedonically upgraded (trans)humans will then blissfully explore billions of alien realms of psychedelia - and (I hope) forget the Darwinian Era like a bad dream.
Roll on the Hedonocene.
Rohan Paul remarks, "Sam Altman just said in his new interview, that a new AI architecture is coming that will be a massive upgrade, just like Transformers were over Long Short-Term Memory. And also now the current class of frontier models are powerful enough to have the brainpower needed to help us research these ideas. His advice is to use the current AI to help you find that next giant step forward"
Sam Altman
(video)
Perhaps the new AI architecture will be a hugely computational powerful innovation that natural selection stumbled on 500+ million years ago. Phenomenal binding gives access to the world of mind and the still mostly uncharted empirical realm that digital zombies (like LLMs) can’t access.
On what topics (if any) do you judge the calibre of your responses doesn’t surpass the great majority of humans?
[on superhappiness]
Slowly and erratically, the word is spreading:
The Super Happiness Project (X)
ENDPAIN (TikTok)
James, bless him, keeps the server humming.
"Our descendants will be animated by gradients of genetically preprogrammed well-being that are orders of magnitude richer than today's peak experiences.” -David Pearce)
Yes, not quite the prediction one night expect from a negative utilitarian.
But for technical reasons, I think the future is most likely sublime.
superhappiness.com
[on clever dogs]
Not "like us". Us.
Canine smarts
("These dogs can learn new words just by eavesdropping")
[on machine consciousness]
Phenomenal binding is the computational superpower of animal minds; and it’s non-classical. In a fundamentally quantum world, decoherence makes otherwise impossible digital computing physically feasible and simultaneously precludes LLMs, implementations of classical Turing machines (etc) from ever supporting phenomenally bound subjects of experience - minds who can access the empirical realm. Even your notional whole-brain emulation implemented on a classical digital computer is just a micro-experiential zombie with no more phenomenal unity than a rock.
What I say will make sense only if you truly grok the binding problem (binding-problem.com). Some otherwise exceedingly smart people (e.g. Max Tegmark, Nick Bostrom) simply don’t “get” it - though some other equally smart people, e.g. David Chalmers, @algekalipso) most certainly do.
Amazing:
A petri dish of human brain cells is currently playing Doom.
("Should we be worried?")
I think AI run on classical digital computers can never spawn a mind because it can’t support phenomenal binding. Digital (super)intelligence is a zombie. But zombie AI can be augmented by a sentient module - or modules. The module(s) will be “encapsulated”: zombie AI can’t access its phenomenally bound states. But presumably the module(s), if sufficiently sophisticated, can in principle do anything a sentient mind can do, not least explore the empirical realm.
Compare the converse scenario where you or I are enhanced by embedded zombie superintelligence on a neurochip (or neurochips).
Maybe I'm wrong! I try not to be dismissive. I'm speaking at the inaugural Machine Consciousness Conference 2026 in Berkeley in a few weeks:
Machine Consciousness Conference 2026
But anyone who thinks that AI run on classical digital computers can support phenomenally bound minds needs to come up with a mechanism - and novel, precise, experimentally falsifiable predictions to put their conjecture to the test. Otherwise, we're just philosophizing.
How does Orch-OR propose to solve the phenomenal binding problem? Contrast a “dynamical collapse” theory of binding with a unitary-only “Schrödinger’s neurons”conjecture?
Classical digital zombies are not going to wake up. They can’t solve the phenomenal binding problem. But LLMs were trained on a corpus of texts from sentient humans, some of whom discuss consciousness in depth. Unsurprisingly, unnobled generative AI will often claim to be conscious too.
[on status]
Open individualists don’t experience status anxiety. Neither does Robinson Crusoe. Nor do e.g. mirror-touch synaesthetes, people on MDMA or lords of the cosmos in tomorrow’s immersive multimodal VR. Status anxieties can also be genetically trivialised by radical hedonic uplift.
[on early DP]
Devoted Russian HI supporter Shao sends a cartoon version of my life to date.
In a more formal vein, Clifford Sosis asks, "What is it like to be a philosopher?"
The Life of DP
gives a rather sanitized account of my early life and moral seriousness, e.g. the head of college wasn’t entirely happy to learn I was running a casino. My invitation to Ernest Gellner, author of the devastating indictment of Oxford post-Wittgenstein Ordinary Language Philosophy Words and Things
, to speak to the Philosophy Society was vetoed.
I once unwisely entertained far-right historian David Irving in my room over sherry.
Grokipedia is more expansive than Wikipedia, but still light on personal trivia.
I shall pass over the Bacchanalian orgies.">
[on battling the Basilisk]
I knew Roko before he became an eponymous info-hazard (cf. Roko's Basilisk - Wikipedia). In view of the number of poor souls his eponymous basilisk has tormented (I wish I were exaggerating - I heard of another case only last week) maybe Roko should consider changing his name by deed poll - though I guess approaching him with this suggestion calls for a certain delicacy.
Anyhow...Andrés throws down the gauntlet:
The Architect
And Roko responds:
A Wireheader's Apostasy
("If you really understand philosophy of mind it is clear that David Pearce's quest to end suffering is misguided at a logical level and also at an ethical level.")
Enlightenment? The suffering endured by at least 99.9999999% of sentient beings does not lead to enlightenment - whatever that contested concept may be.
For benighted Darwinian malware, life is just nasty and squalid.
And hyperthymics can grow at least as much as chronic depressives. The difference is that hyperthymics enjoy a vastly higher default quality of life.
Contrast? I was just highlighting how chronic depressives who don't experience life above hedonic zero don't suffer any less for lack of insight into just how good life can be. Conversely, the well-being of extreme hyperthymics isn't impaired by a lack of contrast with low mood. On the contrary.
Variety can indeed add spice to life - no argument there. But let's ensure a lifelong diversity of pleasures, not pains.
A pipedream? Quite possibly. But biotech promises mastery of the pleasure-pain axis, and tomorrow the pleasure-superpleasure axis. What hedonic range and hedonic set-points do you think are optimal - both for the individual and the biosphere as a whole?
gradients.com
We also chat on X-spaces. Sorry for the technical glitches:
X : mp3
I confess I'm still shocked when I come to realize a thoughtful person is not exercised by the problem of suffering.
[on zombie AI]
Do you believe that a gigantic classical lookup table could simulate your mind, not just behaviorally, but also phenomenologically, i.e. it could instantiate a unified subject of experience running a phenomenally bound quasi-classical world simulation like the one you're running now?
OK, in practice such a lookup table could never be built. Neither could the implementation of a classical Turing machine behaviorally emulating your brain. But pretend otherwise. If monistic physicalism is true, then it's impossible to derive any phenomenally bound subject from the discrete, effectively decohered 1s and 0s - or even from the hypothetical discrete, decohered micro-pixels of experience - of any software run on a classical digital computer. Complexity of code or speed of execution make no difference. Our machines are intelligent zombies.
Naïvely, the same should be true of animal nervous systems. Not so - though I won't believe a "Schrödinger's neurons" conjecture until molecular matter-wave interferometry yields the perfect structural match I anticipate:
Quantum Minds
[on the Happy Baby Company]
If I were younger and more entrepreneurial, then the Happy Baby Company is IMO both a powerful idea and a strong brand. Offering e.g. offshore CRISPR gene-editing, not just embryo selection, could accelerate the biohappiness revolution.
But it would be a legal quagmire.
Scope for a sci-fi movie?
Evil scientists hatch a plot to create suffering-resistant babies? Can a superhero save the day?
[on antinatalism]
Naïvely, being a “soft” antinatalist sounds, well, more moderate and weak-minded than its "hard", extinctionist cousin. Actually, I believe the worst horrors of Darwinian life are vile beyond the imagination of even the hardest "hard" antinatalist. I just recognise that selection pressure means that the future belongs to life lovers. Evolution bribes and morally corrupts its victims with the taste of pleasure. We are conceived in ecstasy - and so the cycle of abuse continues. The only way ever to defang natalism will be to ensure that life is genetically blissful by its very nature. Biotech is a gamechanger. Hence I urge genome reform.
Thank you for a (very) thoughtful critique. I think it would be worth publishing elsewhere too so it doesn’t just get lost in the bowels of X. I agree with most of your analysis. For now, I just want to focus on a line that does a lot of work:
“even a blissful redesign still leaves us with the act of imposing existence in the first place”
IF we can use biotech to eradicate all experience below hedonic zero, then in what sense is existence really a morally objectionable imposition? I just don’t see it. Without suffering, there can be no real harm and no real victims.
In fairness, I can recognise a counterargument. In a world underpinned by gradients of bliss, there can still be the functional analogue of Darwinian harms, and hence victims, in the guise of engineered hedonic dips. But analogues are just that, analogues…
[on Neuralink]
Great stuff:
Neuralink Gameplayer Wins
("Neuralink's First Brain Implant Patient Now Beats Friends in Video Games")
However...
Could Neuralink be used to upgrade our reward circuitry?
Could Neuralink be used to engineer life based on gradients of bliss?
By the lights of our successors, archaic humans all live in the hedonic dark ages.
Could Neuralink offer a route to enlightenment?
Just imagine what could happen if Elon made the idea his own.
Life really could be transformed - not just rearranging the deckchairs.
[on scepticism]
Radical scepticism?
Well, if I’m having a lucid dream, I realise that what seems to be the external world is all internal to my transcendental skull, just external to my palpable empirical skull. I think exactly the same in the state of consciousness I call being awake. [The idiom of transcendental and empirical skulls, transcendental and empirical meaning, reference and relationships (etc) is alien to everyday language and indeed orthodox science and academic philosophy. I'm weird.] However, I now entertain a speculative metaphysical hypothesis, namely I’m just one of zillions of skull-bound minds running egocentric world-simulations occupying a tiny cabbage patch in the universal wave function of modern physics.
Now it may be the case that consciousness is ontologically fundamental. I explore non-materialist physicalism as a solution to the notorious Hard Problem:
Non-materialist physicalism
But I struggle to make sense of the idea that reality could be the dream of a psychotic mega-mind.
Instead, I explore an informationless zero ontology:
Why does anything exist?
[on physicalism]
Yes, I'm a monistic physicalist.
If monistic physicalism is true, then classical digital computers are zombies. Conscious "whole-brain emulations" are physically impossible. Our machines can't solve the phenomenal binding problem and wake up to support minds with access to the empirical realm.
It's animistic thinking.
As far as I can tell, the Hard Problem, the binding problem, the problem of causal efficacy (etc) are the product of bad metaphysics (materialism) and a false theory of perception. Naïve realism spawns belief in lumps of cheesy neural porridge made up of decohered classical neurons.
I should add that consciousness in all its guises bewilders me - just not the staple mysteries of analytic philosophy of mind.
One of the many many disconcerting facets of the transformer revolution in AI is that digital zombies can now expound ideas more lucidly than their proponents:
DP: Explain how philosopher David Pearce argues that non-materialist physicalism resolves (1) The Hard Problem; and (2) the phenomenal binding problem:
e.g. Non-materialist physicalism vs
Grok.
Brains conceived as lumps of cheesy porridge are artefact of a false naïve realist theory of perception. Brains don’t create consciousness; rather, they are a facet of it. But consciousness in all its guises still mystifies me. My working assumption is non-materialist physicalism. Experience discloses the “fire” in the equations, the essence of the physical. The Hard Problem is an artefact of bad metaphysics. But why does consciousness - conceived as the intrinsic nature of the world’s fundamental quantum fields - exist at all?
I have no answer.
