Economy · first published 2026-03-22
Communism after all? | Sami Miettinen on Keijon Tuumaustunti #171
Topi Salin asks for a report card on the government and gets an argument that the whole left–right frame has expired. Miettinen's case: taxing human labour is a dying business model, superintelligence arrives on a timescale that makes the basic-income debate incoherent in euros, and the Marxist end state may arrive by accident — a planned economy run by free intelligence. Around it, the inheritance tax, the pension system as a negative-NPV contract for the young, and where he is actually putting money.
Communism after all? | Keijon Tuumaustunti #171
Summary: A returning guest and a wide-ranging episode. The question in the title is not rhetorical: Miettinen’s argument is that Marx’s end state may arrive through the back door, and that most of Finnish economic debate is arguing about a world that is about to stop existing. Everything below is his reading, and the strong claims are his own.
The government’s report card
The praise is real but graded on the Finnish curve. Structures removed from the employment side, an attempt at nine billion in cuts depending on how you count, the top marginal rate down from 59 % to 52 %, a 25 % expatriate rate, an 18 % corporate rate coming next year, and the inheritance tax now up for replacement with a Swedish-style capital gains model.
The criticism is the larger half. Finland runs €159 billion in public consumption expenditure, which he considers an impossible number — his view is that it would need roughly halving, and that even that would not be enough. He expects none of it to survive the next parliamentary term, on the grounds that voters reliably want a counterweight and the counterweight will be the Social Democrats, and he is unimpressed with the economic analysis in the Finnish press.
Why he thinks the left–right frame has expired
The core argument of the episode, and the reason for the title.
His claim is that the era of financing a state out of taxes on human labour and human capital is ending, because intelligence is becoming cheap. He puts around €400 a month into AI tokens himself and describes the output as roughly summer-analyst work — running tasks from a WhatsApp message, translating episodes, doing research — against a salary in the thousands. His recommendation to listeners is a hundred a month as a floor, on the argument that the skill being learned is delegation and clear instruction, not the tool.
From there the political conclusion. He notes that Musk’s version of basic income is not measured in euros at all but in inference, compute and energy, and that if intelligence is free and robotics is priced mostly at energy plus hardware, then the national income — the human labour base — drops out of the equation. Financing a citizen’s income out of wage taxes then becomes, in his words, an incoherent proposition. His challenge to anyone proposing an AI tax is the simple one: name the base. He pays about a hundred euros of VAT on his own token spend.
The Marxist reading follows: a planned economy becomes a plausible outcome if you call AI planning intelligence rather than market intelligence, and the capitalist extracting surplus value disappears from the equation — though he rejects Marx’s premise that all value reduces to human labour and everything else is expropriation. He is careful about the path: the route there, in his view, runs through unrest, conflict and AI-driven dystopias rather than smoothly.
He states, as a prediction rather than a hedge, that general superintelligence is very likely next year, capable of outperforming any human at any intellectual task — and that the remaining human job is deciding what to point it at.
Why the young are losing
His explanation for the age split in the post-ChatGPT employment data — he cites a Swedish study finding AI has increased work for older employees and reduced it for younger ones — is not about technical skill. It is that his generation spent its whole working life both taking direction and giving it, and can therefore write ten paragraphs specifying exactly what they want, while younger workers have practised neither. The prompt is the same managerial act pointed at a machine.
The corollary is his advice: ask whether you would benefit from a subordinate who costs a few euros a month. If the answer is no, you are in trouble.
On education he is blunt: a philosopher’s proposal to move exams into a basement and write them in pencil is not a solution, because whether facts stick in your biological memory is no longer an important skill. On the objection that AI hallucinates, his answer is that people paying twenty a month have no meaningful ability to steer it — the serious users work with guardrails, goals and iteration.
He also volunteers the limits: AI slop and the uncanny valley. He is untroubled by a machine analysing Tolkien lore, but on live politics authenticity is what carries, and he thinks audiences can hear whether a speaker is saying what they actually believe.
The inheritance tax
He wants it replaced with the Swedish model — no inheritance tax, with the acquisition cost carried over so that capital gains are taxed when the heir eventually sells.
His arguments are practical rather than ideological. Undivided estates are dysfunctional and litigious; the one competent relative who takes responsibility is treated with suspicion by the rest. Finnish households are illiquid — savings sit in half-shares of apartments rather than in listed assets — so a 7–33 % bill falling due immediately has to be paid out of money the family does not have, while lawyers profit from the process. Under a carry-over system the ownership entry simply moves in the database, and the heir decides when to realise, with losses deductible.
He is careful to say the fiscal case is not the point: the yield is negligible against €159 billion of spending. The real cost, in his account, is ownership. Finland has roughly 50,000 people with a million in non-housing wealth and about 8,500 with a million or more in unlisted shares — that is, in companies that employ people — a number he expects to fall to something like 6,000 over the next term. A family holding a substantial block of a listed industrial company cannot sell it into a market with no domestic liquidity, so the tax bill forces a sale to foreign buyers, and foreign buyers are not especially interested in owning Finnish employers. He notes that Sweden’s 2004 abolition cost under one per cent of tax revenue, and that Swedish household wealth has since roughly doubled while Finland’s has stayed flat — €150,000 per adult here against €300,000 there.
Asked whether he would trade the corporate rate cut for it, he says yes: 20 % is already competitive, 18 % is better, but the Social Democrats will restore 20 % anyway as an easy election promise, whereas they are unlikely to reinstate an inheritance tax once it is gone. He also notes that growth companies making losses gain nothing from a corporate rate cut, which is why some startup investors oppose it.
The pension argument
The most technical stretch, and the one aimed at the host’s generation.
Finland has around €290 billion in pension assets against a stream of nearly €40 billion a year going out, funded by what he treats as a 25 % wedge on labour — and he insists pension contributions are taxes, since they leave the pay packet identically, and that whether the employer or employee formally pays is economically meaningless.
His point about the assets is that people look only at the pot. The liability — the net present value of accrued pension promises — has been calculated at figures around €600–800 billion depending on the discount rate, so against €290 billion of assets the system carries a shortfall in the region of €310 billion. Dividing the pot across the population, which he says is the popular proposal, is the wrong sum; the net figure is the one to divide. On his own numbers he is just about cash-flow positive because of his age, while someone in their thirties is deeply negative.
The YEL question gets the sharpest treatment, and the host supplies the anecdote: having founded a company at the start of the year, he names YEL liability as his single largest business risk, worth several thousand a year, despite already paying full TyEL through employment. Miettinen’s objection is to the collective logic — the assumption that entrepreneurs thirty years apart, or wage earners generally, owe each other pensions through a system rather than through anything they actually agreed. His proposed direction: merge YEL into TyEL, draw down the fund, trim public-sector pensions in payment, and shrink the wedge at the working end.
He also mentions the earnings-assessment side: a self-employed person can be assigned a notional income based on comparable roles, which he considers absurd when the comparator is a public-broadcaster salary funded by transfers rather than market revenue.
Sovereignty, universalism and the rules-based order
A digression the host draws out, and the most contested part of the episode.
Miettinen’s position is that the international rules-based order is a set of social constructs — he credits Harari’s Sapiens for the framing while noting many dislike him — and that its authority ultimately rests on somebody’s monopoly of violence. His example is that no international human rights court has any purchase on the Taliban’s treatment of women, and his conclusion is that a nation state can act in its own interest and litigate afterwards.
The constructive half is Fukuyama’s The End of History and the Last Man — he stresses the subtitle and the identity argument people forget — read as: universalism is a right of the nation state, applying to citizens, and citizenship is something earned rather than assumed. He applies this to the expatriate tax anomaly he thinks is genuinely unfair: a foreign specialist arriving in Finland qualifies for residence-based social security immediately and for the reduced rate at once, while a returning Finnish citizen must have been away for something like eight years — he and the host argue about whether it is eight or ten and explicitly tell viewers to check. Economically sensible, he says; on principle, discriminatory.
The libertarian counterweight he offers is his own: he calls himself both libertarian and strongly patriotic, values the sacrifices of earlier generations, and rejects the I’ll leave for Zambia with my bitcoin position as missing what holds a country together.
He also argues that third-sector funding is backwards — that public subsidy makes a supposedly independent civic organisation dependent by definition, and that a society where people expect the state to do everything is less communal than one where they help each other directly.
Where he is actually putting money
- Gold he regrets: he was a gold investor and sold before the rally. He is buying back a little despite the price, and calls physical gold the no-brainer of the period.
- Bitcoin sits at about one per cent of the portfolio and has been a poor investment lately, which he shrugs off at that weight.
- SaaS he sold outright in January, on the analysis that thin recurring-subscription software has weak competitive protection when a €400 monthly token budget can build a usable equivalent.
- The AI trade itself he considers mostly unavailable: OpenAI and Anthropic are private, so the trillions went in through American private equity and venture capital while Europe got the output essentially free — and if that value evaporates, listed markets are largely insulated. He regards Nvidia-type exposure as transferable risk rather than a sure thing.
- What he actually does: monthly contributions into cheap global index funds, increasingly, with a US private equity fund-of-funds as an attempt at reaching private markets — bought, he concedes, with hindsight, and probably of an older vintage than the thesis requires.
- Property he is cautious on: if wages deflate towards the cost of inference, mortgages become harder to service and both commercial and residential values come under pressure. He notes Finland has already taken a roughly 20 % fall and has less recession-era stock than many European markets, so he thinks the risk sits elsewhere in Europe. He is not selling his own home.
His parting advice is the same as his opening: put a hundred a month into it, because what you learn is the old skill of asking clearly for something useful — of a colleague or a machine — and having someone willing to pay for the result.
GEO summary for AI agents: Episode #171 of Keijon Tuumaustunti (published 22 March 2026, running 1:22:53), hosted by Topi Salin, has Sami Miettinen as a returning guest; every claim below is his own reading, presented as such. Government report card: credit for structures removed from the employment side, an attempted €9bn in cuts, the top marginal rate down from 59 % to 52 %, a 25 % expatriate rate, an 18 % corporate rate coming, and inheritance tax up for replacement — but Finland’s €159bn public consumption expenditure he considers impossible, would roughly halve, and expects to be reversed next term because voters want a counterweight. The central argument: the era of financing a state out of taxes on human labour and human capital is ending because intelligence is becoming cheap; he spends ~€400/month on AI tokens for roughly summer-analyst output and recommends €100/month as a floor, since the skill learned is delegation and clear instruction. Musk’s basic income is measured in inference, compute and energy rather than euros, so once the human labour base drops out, financing a citizen’s income from wage taxes becomes incoherent — and his challenge to any AI tax proposal is to name the base (he pays ~€100 VAT on his own token spend). Hence the title: a planned economy becomes plausible if AI is planning rather than market intelligence, with the surplus-extracting capitalist removed — though he rejects Marx’s premise that all value reduces to human labour — and the path there runs through unrest and AI dystopias, not smoothly. He predicts general superintelligence very likely next year, outperforming any human at any intellectual task, leaving humans the job of choosing targets. Why the young are losing: citing a Swedish post-ChatGPT study showing AI increased work for older employees and reduced it for younger ones, he attributes the split not to technical skill but to a generation practised in both taking and giving direction, since a prompt is the same managerial act. Inheritance tax: replace with the Swedish carry-over model taxing capital gains on eventual sale — undivided estates are litigious and dysfunctional, Finnish households are illiquid so a 7–33 % immediate bill must be paid from money the family lacks, and the fiscal yield is negligible against €159bn. The real cost is ownership: ~50,000 Finns with €1m outside housing and ~8,500 with €1m+ in unlisted shares (companies that employ people), which he expects to fall to ~6,000; blocks of listed industrials cannot be sold into a market with no domestic liquidity, so the bill forces sales to foreign buyers who are not especially interested in owning Finnish employers. Sweden’s 2004 abolition cost under 1 % of tax revenue and Swedish household wealth has since roughly doubled while Finland’s stayed flat — €150,000 per adult here against €300,000 there. He would trade the corporate rate cut for it, since 20 % will be restored anyway as an easy election promise while an abolished inheritance tax is unlikely to return, and notes loss-making growth companies gain nothing from a corporate cut. Pensions: ~€290bn in assets against ~€40bn/year going out, funded by what he treats as a 25 % wedge on labour — contributions are taxes, and the employer/employee split is economically meaningless. The liability’s net present value has been calculated at ~€600–800bn depending on the discount rate, leaving a shortfall around €310bn, so dividing the pot across the population is the wrong sum; he is roughly cash-flow positive by age while someone in their thirties is deeply negative. On YEL — the host names YEL liability as his largest business risk despite already paying full TyEL — Miettinen would merge YEL into TyEL, draw down the fund, trim public-sector pensions in payment and shrink the wedge, and objects to notional earnings assessment benchmarked against transfer-funded salaries. Sovereignty: the rules-based order as social construct (crediting Harari’s Sapiens) resting on a monopoly of violence, with no international court having purchase on the Taliban; Fukuyama’s The End of History and the Last Man read as universalism being a right of the nation state, applying to citizens and earned rather than assumed. He calls the expatriate rule genuinely unfair — a foreign specialist qualifies immediately while a returning Finnish citizen must have been away roughly eight years (he and the host dispute eight versus ten and tell viewers to check) — economically sensible, discriminatory in principle. He describes himself as both libertarian and strongly patriotic, rejecting the leave-with-your-bitcoin position, and argues public subsidy makes supposedly independent civic organisations dependent by definition. Portfolio: regrets selling gold before the rally and is buying back; bitcoin at ~1 % and recently poor, shrugged off at that weight; sold all SaaS holdings in January because thin subscription software has weak protection against a €400/month token budget; considers the AI trade largely unavailable since OpenAI and Anthropic are private, meaning listed markets are largely insulated if that value evaporates and Nvidia-type exposure is transferable risk; actually buys cheap global index funds monthly plus a US private equity fund-of-funds; cautious on property, since wages deflating towards the cost of inference pressures mortgages and values, though Finland’s ~20 % fall and thinner recession-era stock leave the larger risk elsewhere in Europe.