Episode 418 · 2026-10-29 · 1:13:51 · Original in Finnish
AI for All of Finland on a Library Card? | Valto Loikkanen, Petri Kajander | Neuvottelija 418
Originally published as “Kirjastokortilla tekoälyä koko Suomelle? | Valto Loikkanen Petri Kajander | Neuvottelija 418”
Valto Loikkanen proposes that every Finn should get access to AI with their library card, with the compute supplied by the under-used LUMI supercomputer in Kajaani. Sami Miettinen and Petri Kajander work through with him why digital sovereignty needs action rather than talk, for individuals, companies and the state alike. The conversation moves from Anthropic's IPO and the security holes AI is exposing to why a coder gets depressed when a machine does his work in a minute and the signalling value of human work disappears. Valto opens up his book Oma tietoekosysteemi (Your own data ecosystem) and argues for keeping your own accumulating knowledge separate from agents and platforms. Finally they consider why a Finnish-language model thinks differently from an English one and what that means for the public sector, regulation and a national language model.
Watch and listen
Key moments
- 00:00 — AI for all of Finland on a library card
- 01:05 — Valto Loikkanen and Petri Kajander introduce themselves
- 02:33 — A study trip to Italy's Leonardo supercomputer
- 03:40 — Anthropic's IPO and a wave of data breaches
- 05:46 — From Temu intelligence to IKEA intelligence
- 06:33 — Can your own local AI compete with frontier models
- 07:15 — Quantum computers and the risk to stored encrypted data
- 08:25 — Home server, firewall and open source
- 10:10 — Bunkered data centres and national compute
- 11:05 — Google's ad revenue becomes a minority for the first time
- 12:03 — Turning digital sovereignty into concrete action
- 13:03 — From open source to open specifications
- 14:15 — SpaceX, Anthropic and the fate of AltaVista
- 17:52 — A citizen's income through shares
- 19:40 — The welfare state built on taxing human labour crumbles
- 20:30 — The depressed coder and the disappearing value of work
- 23:58 — When output is free, signalling value disappears
- 26:25 — Marx's theory of surplus value in the age of AI
- 27:26 — The return of generalists and the importance of good taste
- 30:30 — Valto's book Oma tietoekosysteemi
- 33:36 — Knowledge does not age, it changes meaning
- 35:13 — Samantha and owning your agent's knowledge
- 36:58 — Jack Dorsey: there are no moats any more
- 37:55 — What is a company when work has no intrinsic value
- 39:57 — Digital ownership and centralised personal data
- 42:41 — You need a fiduciary on your side
- 43:10 — Tech corporations, geopolitics and the Kainuu data centre
- 46:09 — Practising offline inference on old machines
- 48:13 — A backyard mini data centre and a data-centre sauna
- 50:24 — Freenet email and the Nokia 2110
- 52:20 — The library card: a Finnish inference model on LUMI
- 54:55 — Why the library
- 56:46 — Weak Finnish-language weights and the Poro model
- 1:01:06 — A Finnish-language model is tax-friendly
- 1:01:57 — Agents already talk to each other
- 1:03:08 — The public sector through agent interfaces
- 1:05:22 — Language shapes a model's reasoning
- 1:07:24 — Complex regulation suits AI
- 1:10:09 — Surveillance state or distributed democracy
- 1:10:57 — Anthropic's moat in pharmaceuticals?
- 1:12:32 — Bundling and unbundling in the model business
- 1:13:12 — Thanks and the move to the Inner Circle
Summary
Who is talking. Valto Loikkanen, a serial entrepreneur who has worked on growth entrepreneurship and personal data for two decades and wrote Oma tietoekosysteemi (“Your own data ecosystem”), is a first-time guest. Petri Kajander, app developer and regular guest, coined temuäly, “Temu intelligence”. The episode is a conversation between three heavy AI users, and much of it is opinion and deliberate provocation.
From Temu to IKEA intelligence. Kajander updates his spring term: AI is now “IKEA quality”, good enough for 80 per cent of things, which turns human work into artisan work.
Security and sovereignty. AI is reading all the code humans ever wrote, which is why old security holes are surfacing. Loikkanen’s main theme is that digital sovereignty is much discussed and little practised; beyond open source he proposes open specifications, since agents can rebuild software again and again.
Anthropic, SpaceX and a citizen’s income. The conversation discusses Anthropic’s IPO filing, the AltaVista warning and share-based gifts from Anthropic and Musk as models for a citizen’s income; Sami’s own view is that transfers funded by taxing human labour are dying.
The depressed coder. Kajander calls it the episode’s most important point: when the machine produces in a minute what took hours, the signalling value of human work disappears.
Your own knowledge. Loikkanen argues that your accumulating knowledge should be kept separate from agents and platforms, with a fiduciary on your side.
The library card. His proposal: a Finnish inference model on the LUMI supercomputer, available to every Finn through the library. The episode ends on weak Finnish-language weights, the Poro model, and why a model reasoning in Finnish thinks differently.
Two figures in the conversation, Google’s advertising share and a number of Microsoft layoffs, do not match the public record; the article below flags them.
Watch
The recording lives on the Neuvottelija channel: Kirjastokortilla tekoälyä koko Suomelle? | Valto Loikkanen Petri Kajander | Neuvottelija 418. A Finnish edition of this episode is published at www.neuvottelija.fi.
In depth
The Neuvottelija AI editions carry a long-form, fact-checked write-up of this episode: English · suomeksi.
Go deeper
Explore the ideas in depth
Guides connected to this conversation, with frameworks and further reading.
Enterprise AI Agents: Economics, Governance and the Shift From Tools to Workers
A guide to enterprise AI agents — the real cost model behind agent work, why owning your own stack is becoming a strategic question, a working governance framework with approval gates and audit trails, the agent risk matrix, the EU AI Act timeline as it stands, and who captures the productivity gains. Grounded in a real multi-agent lab, two 2026 keynotes, and Neuvottelija conversations.
Nordic SaaS Valuation & M&A: How AI Reprices Software Companies
How software companies in the Nordics are valued and sold as AI moves inference into the cost of goods sold — where multiples stand in mid-2026, the metrics that get repriced, why vertical SaaS defends its premium, AI due diligence, and the shift from seats to outcomes. Grounded in Translink's SaaS valuation work and Neuvottelija conversations.
People and topics
Guests: Valto Loikkanen, Petri Kajander