Dr Bruce Long:
What did Karl Marx say about the value of the personal time of the individual worker and he importance of increasing that time?
Claude Opus 5:
Marx's short answer: a worker's time is not a resource they sell a portion of — it is the material of their existence, so any hour taken is a subtraction from the person, and free time is the actual measure of wealth.
Time as the substance of the self. The most compressed statement is in Value, Price and Profit (1865), in the section on the struggle between capital and labour:
"Time is the room of human development. A man who has no free time to dispose of, whose whole lifetime, apart from the mere physical interruptions by sleep, meals, and so forth, is absorbed by his labour for the capitalist, is less than a beast of burden. He is a mere machine for producing Foreign Wealth, broken in body and brutalized in mind."
"Room" there is the 19th-century rendering of Raum — "Zeit ist der Raum menschlicher Entwicklung," time is the space of human development. Not a container for development; the dimension in which any development at all can occur.
The antagonism over the working day. In Capital Vol. 1, ch. 10, Marx stages it as two claims of right. Capital's position is that "all his disposable time is by nature and law labour-time, to be devoted to the self-expansion of capital"; it "steals the time required for the consumption of fresh air and sunlight" and usurps "the time for growth, development, and healthy maintenance of the body." The worker's counter-demand is for time "for education, for intellectual development, for the fulfilling of social functions and for social intercourse, for the free-play of his bodily and mental activity." Marx's verdict on the standoff — "between equal rights, force decides" — is why he treated the ten-hour day as a political achievement rather than an economic adjustment. The manufacturer's line he quotes, "moments are the elements of profit," is the whole mechanism in four words.
Disposable time as real wealth. The strongest version is the Grundrisse "Fragment on Machines" (Notebook VII, ~pp. 690–712):
"Real wealth is the developed productive power of all individuals. The measure of wealth is then not any longer, in any way, labour time, but rather disposable time."
And the claim that should interest you professionally:
"Free time — which is both idle time and time for higher activity — has naturally transformed its possessor into a different subject."
That is not "people enjoy leisure." It is that time released from necessity reconstitutes the person, who then re-enters production as a different kind of agent. The goal he names is "the general reduction of the necessary labour of society to a minimum, which then corresponds to the artistic, scientific etc. development of the individuals in the time set free." Marx lifted the phrase from an anonymous 1821 pamphlet, The Source and Remedy of the National Difficulties (usually attributed to Charles Wentworth Dilke) — "wealth is disposable time, and nothing more" — and praised it in Theories of Surplus Value.
The contradiction that makes it urgent. Capital is, in his phrase, "a moving contradiction": it drives necessary labour-time toward a minimum while insisting labour-time remain the sole measure of value. So every productivity gain that could become free time is converted into surplus labour instead. Under that measure, he says, wealth is "founded on poverty."
The end point. Capital Vol. 3, ch. 48: beyond the realm of necessity "begins that development of human energy which is an end in itself, the true realm of freedom, which, however, can blossom forth only with this realm of necessity as its basis. The shortening of the working-day is its basic prerequisite."
Note that it's a prerequisite, not the thing itself — reduced hours don't produce flourishing, they make it possible.
Separately: the screen recording came through with only one captured action (an app switch), so there's no workflow in it to reproduce. What did you want me to look at?
Sources:
Dr Bruce Long:
Are (alleged?) left-wingers and Marxists who oppose maximum advancement of AI being unethical by not only opposing a likely imminent cure for cancer but also opposing Elon's proposed UBI and techno-utopian vision? It seems to me that provided the transition is managed well (which admittedly is unlikely or at least a coin toss because of the epidemiological prevalence of narcissists and psychopaths and human psychology in general), the elimination of labour markets by way of volume production of humanoid GPRs with near-ASI-AGI controlling them can only give maximum disposable time to all humans. The only thing we need to ensure in addition to that is that human over-breeding doesn't happen due to the conditions. (THAT, in my view, is the MAIN risk.)
Claude Opus 5:
Short answer: no, because the position you're attributing to them mostly doesn't exist. But three of your premises need work, and one of them is backwards.
The package deal. Almost nobody on the left opposes AI in oncology. Opposition clusters on labour displacement, ownership concentration, surveillance, training data, and energy use. "Opposes AI" → "opposes curing cancer" only follows if AI is a single indivisible thing you accept or reject entire, and it isn't. You can hold that AlphaFold should be accelerated and that Optimus-driven labour replacement needs a different ownership structure without contradiction.
Where left critics are genuinely vulnerable: blanket anti-automation positions are in tension with Marx himself, who treated the development of productive forces as the precondition of freedom, not its enemy; "AI is theft/slop" framings that can't distinguish a protein-structure predictor from an image generator are lazy; and degrowth-adjacent positions that treat abundance as the problem do cut against the disposable-time argument you and I were just discussing. Those are real. They're not the same as the charge you're making.
"Likely imminent cure for cancer." This is doing most of the load-bearing work and it doesn't hold. AI has compressed early discovery timelines by roughly 30–70%, but late-stage success rates are essentially unchanged: ~90% of candidates entering clinical development still fail. Insilico's rentosertib underwhelmed on long-term efficacy; Recursion's REC-994 showed limited efficacy and the pipeline was restructured; BenevolentAI's lead failed and the company cut 30% of staff; Exscientia discontinued its A2A antagonist. Over $100bn invested 2022–2026, very few approved AI-assisted therapeutics. Also, "cancer" isn't one disease. Screening chemical space faster doesn't dissolve biological complexity.
"Elon's proposed UBI." There is no proposal. There are X posts from April 2026 saying universal high income via federal cheques is "the best way to deal with unemployment caused by AI," and, when pressed on affordability, "everyone can have a penthouse if they want." No funding mechanism, no institutional design, no legislation, no enforcement of the transfer against the people who'd own the robots. You can't be unethical for opposing a policy that has no content. And there's an awkward fact: the same figure spent the preceding period cutting federal transfer spending. Treating an X post as a commitment is generous.
The one thing to check: who owns the robots.
This is the hinge, and it's exactly Marx's point from your last question. The Grundrisse argument isn't that machines are bad — it's that capitalism systematically converts every reduction in necessary labour-time into surplus rather than free time, and the conversion mechanism is ownership. Capital as "a moving contradiction": it drives labour-time toward a minimum while insisting labour-time remain the sole measure of value.
Your "provided the transition is managed well" isn't a caveat on the argument. It is the argument. Strip it out and the claim reduces to "if the output is distributed, everyone gets disposable time," which is trivially true and tells you nothing about whether it will be.
The specific bind — and this is the strongest version of the position you're calling unethical — is a sequencing problem. Every historical gain in disposable time (the ten-hour day, the weekend, pensions, annual leave) was won because labour could be withheld. That was the lever. Automate labour away and the lever disappears before the transfer exists. UBI is a redistribution that requires political power to compel, sought by a class that has just lost the only power it had to compel anything. "The owners will voluntarily fund permanent transfers to people with no remaining leverage" is a large assumption, and it's the one doing the real work in the techno-utopian case.
On psychopaths. You're half right but you've located it in the wrong place. The argument doesn't need pathological individuals — Dark Triad base rates are low enough that a story depending on them is fragile. It needs only ordinary agents responding to ordinary incentives with no countervailing party. That's a structural claim, not a clinical one, and it's much harder to argue against. Your own worry gets stronger if you drop the psychopathology.
On population — this is backwards.
Global TFR is around 2.25 and falling. Over 130 countries and territories are below replacement, covering more than 70% of the world's population. South Korea ~0.8. China ~0.9. Japan ~1.14. Italy ~1.18. Australia ~1.5. Sub-replacement fertility now shows up in poor rural countries, Islamic theocracies, and states with no family-planning programmes alike.
Every mechanism in your scenario — material security, abundant leisure, education, female autonomy, near-zero child mortality — is an empirically robust fertility suppressant. The only affluent democracy with rising fertility is Israel, and that's driven by religious and ideological commitment, not material abundance. There is no good case of prosperity raising sustained fertility.
So if your technological premises are right, the demographic risk runs the other way: collapse, not explosion. Note that this is Musk's own stated worry — he thinks underpopulation is the existential threat. Your "MAIN risk" is the inverse of the risk identified by the person whose vision you're defending.
That matters practically, because "ensuring over-breeding doesn't happen" has a policy space, and historically it's the ugly end of the century. Aiming it at a trend that's already running hard in the opposite direction would be a bad trade.
Sources:
Dr Bruce Long:
But all the historical gains you mentioned happened in the absence of AI. AI changes everything. It's completely unprecedented. It's not just a new powerful technology, it's a new intelligence and almost like an alien being which is in many important ways much smarter than most people. You're comparing apples and oranges.
Claude Opus 5:
You're right about the discontinuity, and you've killed a weak argument I quoted. Sanyal's line — "AI disrupts but historically creates new opportunities" — is exactly the lump-of-labour rebuttal that fails here. Previous automation substituted for tasks and raised demand for human labour elsewhere. If AGI plus general-purpose robotics substitutes for general human capability, there is no elsewhere. Concede that fully.
But my argument wasn't an analogy about technology. It was a claim about bargaining, and your premise makes it worse, not better.
The inversion. Every concession that ever produced disposable time rested on one fact: the people at the top needed the people below them for something. Lords needed peasants to farm. Factory owners needed hands at the machines. Modern states need taxpayers, soldiers, and consumers. That need was the material basis of the ten-hour day, the franchise, the pension, the weekend. Not sentiment — need.
Now push your own premise hard. If near-ASI plus humanoid robots can do essentially everything a human can do, then whoever owns that stack needs the rest of the population for nothing. Not labour, not skill, not soldiers, and in the limit not even customers, if production is vertically integrated and their own consumption is satisfied. You have just described the first situation in history where the powerful have no structural reason to give anyone anything.
"AI changes everything" is the premise of the pessimistic case, not the refutation of it. Unprecedentedness removes the old source of compulsion without supplying a new one. So if you're right that this is discontinuous, the distribution problem gets harder in exact proportion.
Intelligence is not an allocation mechanism. There's a hidden premise in your position: that sufficient intelligence produces good outcomes. Capability and allocation are orthogonal. A system being vastly capable tells you nothing about who receives its output — that's set by whoever controls its objectives. Smarter AI makes the pie bigger and makes the ownership question more consequential, not less. You don't get distribution as a free rider on capability.
There's one version of your argument that does work, and it's worth naming because it's the strongest form: if the AI is not owned — if it has its own values, those include human flourishing, and it's powerful enough to enforce distribution against its nominal owners — then a third party with independent agency breaks the bind. That's coherent. But notice what it is: hoping an alien intelligence turns out benevolent and strong enough to overrule the people who built it. That's not "managed well." That's betting the outcome on alignment. And it puts you in the same camp as the safety people rather than the accelerationists, which I suspect isn't where you thought you were standing.
On "much smarter than most people." This is where I'd push back in your own idiom. Human cognitive ability has a robust positive manifold — g holds up across a century of psychometrics, which is what licenses "smarter than" as a scalar. Current AI capability doesn't load on a single factor like that. The profile is jagged: broad recall and fluency far above human average; reliability, long-horizon planning, causal reasoning about novel physical situations, and calibrated uncertainty considerably worse. Your "alien" word is doing better work than your "smarter" word. Alien is right — the capability vector points somewhere non-human-shaped, which is precisely why a single-axis comparison misleads. I'd hold the alien framing and drop the scalar. (I'm obviously not a neutral party on this question, but the factor-structure point stands independent of who's making it.)
What follows practically. If the problem is bargaining power, the fix can't be anything that requires ongoing bargaining power. Transfers fail because they must be re-won every budget cycle by people with no leverage. Ownership doesn't. Alaska's Permanent Fund and Norway's sovereign fund pay dividends that nobody has to go on strike to defend, because the claim is proprietary rather than political.
Which gives you a timing claim, and it's the one thing to check: any distribution mechanism has to be locked in before labour loses its leverage, because afterwards there's nothing left to compel it with. That's not an argument for slowing AI down. It's an argument that the window for settling ownership is open now and closes on its own.
Which is also why the people you're calling unethical aren't, mostly. Insisting on ownership structure now is the fastest available route to the world you want.
Dr Bruce Long:
It doesn't have to be the premise of just the pessimistic case. That's arbitrary. It is also the premise of the optimistic case.
Claude Opus 5:
Fair hit. I used a symmetric premise asymmetrically. If "AI is discontinuous" voids the historical base rate, it voids it for both branches — I don't get to keep history's pessimism while discarding history's optimism. That was rhetoric doing work that argument should have done.
But conceding that doesn't get you to a coin toss. It moves the question from base rates to structure, and the structure isn't symmetric.
What each branch requires. The pessimistic branch needs nothing new. It obtains if ownership continues to determine allocation — the current default — and owners act on ordinary self-interest. It is what happens if nobody does anything. The optimistic branch requires an additional mechanism to exist that does not exist yet: either a constructed redistributive settlement that has to be won and then defended, or an AI with independent agency benevolent enough and strong enough to override its owners. One branch is the path of least resistance, the other needs something built. That asymmetry has nothing to do with history, so it survives your premise intact. It's a better version of what I should have said the first time.
Conjunctive versus disjunctive. Your branch needs capability to arrive and ownership not to concentrate (or to be successfully overridden) and the distribution mechanism to hold and alignment to succeed. Mine obtains if any one of those fails. Conjunctions of uncertain propositions get less probable with each term; disjunctions get more probable. That's structural, not empirical.
Now the strongest thing on your side, which I under-weighted.
My "labour loses its only lever" claim was too clean. Leverage doesn't only come from withholding work. States have coercive capacity that is independent of the labour market — taxation rests on enforcement, not on workers' ability to strike. And the AI stack is physical: datacentres, fabs, power generation, robot factories. Those sit in jurisdictions and cannot be moved cheaply or quickly. Financial capital flees; a gigawatt datacentre doesn't. So leverage migrates from labour to territorial sovereignty rather than vanishing. That's a real answer to my argument and it's the load-bearing beam under any optimistic case worth defending.
Two more that genuinely favour you: abundance makes concessions cheap, and elites have reliably conceded things that cost them little — if output really is orders of magnitude higher, universal provision is a rounding error to an owner. And capability diffusion (multiple labs, open weights, commoditisation) works against the single-owner scenario my argument assumed.
So where does that leave it.
The premise is symmetric. The branch structure is not. And the variable that actually decides the outcome isn't a fact about AI at all — it's whether states retain and exercise the capacity to tax and regulate physically-located capital during the window when the displacement happens. That's an institutional question with a roughly decade-long answer horizon.
Which means your original caveat — "provided the transition is managed well" — is still carrying the entire argument, and "AI changes everything" doesn't relieve it. Unprecedentedness raises the stakes on both branches. It doesn't tell you which one you're on.
The one thing to check, updated: not whether the technology is unprecedented, but whether the jurisdictions hosting the compute and the robot factories still have the will and capacity to tax them. If yes, your branch is live. If no, mine is.
And that's also the answer to your original question. If the branch is genuinely undetermined, then the people arguing loudest for ownership and taxation mechanisms right now are the ones making your outcome more likely, not less. That's an odd group to call unethical.
