guide to agentprivacy
Browse collections
โœจVisualise
Connect with Star
Your VTA, your chosen perspective

The planned connection uses your VTA and the Trust Spanning Protocol to carry a scoped exchange for you or your agent. You choose what is presented; the receiving service checks the request before a view is shared.

This guide has no VTA connection adapter yet. Opening Star does not connect an identity or send a key.

Open Star โ†— ยท Inspect your City Key โ†—
guide / Vision / Sovereign Compute

A strategy conversation that runs the Sovereign Compute session through the team's standard discovery questions (what is it / who cares / who is involved / what changes / what is the proof / what makes it real).

HTML5 mp4 http://david.local/assets/workshop-videos/sovereign-compute.mov
Discovery session running the "Living Sovereign Compute" milestone through the who-cares / what-has-changed / proof questions for the Hitchhiker / Game of 42 project.

A structured discovery session that runs the "Living Sovereign Compute" milestone through the team's standard interrogation questions โ€” who cares, who is involved, what has changed, what is the proof, and what makes it real โ€” for the Hitchhiker / Game of 42 project. The premise is a stretch-goal achievement: a peer-to-peer network of sovereign data centres and home labs, networked together with national compute and data capacity, that is economically self-sustaining, growing, and credibly on a path to out-computing the big foundation models. The group maps the constituency in both a negative and a positive sense โ€” centralised compute providers and the frontier models on one side (with an awkward timing conflict: hard to partner with Anthropic and OpenAI on Deep Thought while launching an alternative and asking for ~42 billion in compute); the data-dignity movement, indigenous communities wanting culturally-owned language models, governments (EU, UK, anyone apart from American/Chinese), home labs, universities, libraries, satellite-compute players, and hardware vendors on the other. The recovered second half pulls in the open-source AI world โ€” Stable Diffusion and Emad Mostaque's departure from Stability AI in March 2024 toward governance, his SAGE (Sovereign AI Governance Engine) inside the Intelligent Internet project โ€” plus thinkers like Ilya Sutskever (Safe Superintelligence), Andrej Karpathy, Mustafa Suleyman, the personal-AI "choir" community, Nous Research, Mistral, Holochain compact-datacentre builders, NVIDIA/Microsoft and Apple. A recurring strategic thread reconciles the apparent conflict between provocatively launching sovereign compute and courting frontier labs, resolving it via the Hitchhiker framing where citizen assemblies and a "trust ceremony" gate any foundation-model involvement before reaching Deep Thought.

Questions

Key points

  • The milestone defined: Living Sovereign Compute = a viable peer-to-peer network of sovereign data centres / home labs plus national compute and data capacity, economically sustainable, growing, and credibly able to outpace the big foundation models. The session is run through the standard discovery questions (who cares / who is involved / what has changed / proof / becoming real).
  • Who cares (negative / complicated): centralised compute providers and smaller shop-style data centres seeking dynamic compute distribution; the frontier models, who care in a negative sense given the Deep Thought partnership timing conflict (launching an Anthropic/OpenAI alternative while asking for ~42 billion in compute); China in a funny way given its open-source-model USP. Anthropic is framed as wanting a balance โ€” engaging the community on the risks of unguarded, unregulated open-source AI to be seen as a good player.
  • Who cares (positive): the data-dignity movement (self-sovereign control of one's data footprint), indigenous groups wanting models trained on their own culture and language, EU/UK and non-US/non-Chinese governments, universities and researchers who don't want compute subject to the whims of the big models (post-Fable), the UN, open-science advocates, satellite-compute and decentralized-data plays, and libraries as funded local-compute community hubs.
  • Open-source AI roll-call (recovered segment): Stable Diffusion / Stability AI and founder-former-CEO Emad Mostaque (left Stability March 2024, now focused on governance), his SAGE โ€” Sovereign AI Governance Engine inside the Intelligent Internet project (open-source, agent-driven policy platform using AI agents, policy simulators, sovereign compute and multilingual co-pilots to help states design, test and adapt governance frameworks); plus Ilya Sutskever (Safe Superintelligence), Andrej Karpathy, Mustafa Suleyman, Nous Research, Mistral, Midjourney/LLaMA as reference points, the personal-AI choir community, Holochain compact-datacentre builders, and Audrey Tang.
  • The Hitchhiker / Game of 42 strategy: uses the Hitchhiker's Guide cast as a frame โ€” Eddie the shipboard computer and local-software/compute characters vs. Deep Thought. Present both sides: strong arguments for local AI sovereignty and its risks, and the case for carefully-audited cloud-only compute and its risks โ€” because we're all about perspectives.
  • The timeline resolves the conflict: a proliferation of citizen assemblies lets the project tell frontier labs (e.g. Anthropic) they must participate in deliberation and a trust ceremony before reaching Deep Thought / contributing their tokens, so the provocative decentralized launch and the partnership ambition coexist.
  • What has changed: isolated home labs (Australia, Berlin, elsewhere) gain practical grounding to combine forces and feel big enough to be heard; a real feedback loop where data contributors start earning real money; a mindset shift toward privacy and per-person data ownership at the local hub; libraries reimagined as funded local-compute public services; the ability to run big global events on self-run compute rather than the cloud. NVIDIA (DGX / Spark) with Microsoft, plus Apple, care because it opens a new local-AI hardware market.
  • The launch event: an initiative pegged to November 5th, 2027, set against the disaster scenario of centralized foundation compute, mobilising the same constituency โ€” artists, podcasters, thinkers, open-source-AI companies, home labs, decentralized-data plays, and government officials.
  • Proof & realness: the core proof is the number of nodes and the total amount of addressable data, plus event turnout; hard KPIs/numbers are deferred (a little too early), to be set when the milestone is scheduled (expected within the next few months) and to become the project plan.

Action items

  • Define how the sovereign-compute network interfaces with existing data-dignity / decentralized efforts already underway (flagged as a follow-up question).
  • Use the citizen-assembly + trust ceremony mechanism as the gate for frontier-lab (Anthropic) involvement before any Deep Thought / token contribution.
  • Resolve the Deep Thought partnership timing conflict before publicly positioning against the frontier models / the ~42-billion compute ask.
  • Present both sides in materials โ€” strong arguments and risks for local AI sovereignty, and for audited cloud-only compute โ€” keeping the all-about-perspectives framing.
  • Engage the open-source-AI constituency: reach Emad Mostaque / SAGE / Intelligent Internet, Nous Research, Mistral, the personal-AI choir community, and Holochain compact-datacentre builders (Hesu/Heikey Cabrera + the Dutch datacentre-manufacturing builder).
  • Approach hardware vendors opening the local-AI market โ€” NVIDIA (DGX / Spark) + Microsoft, and Apple โ€” as interested who-cares partners.
  • Pitch libraries and local councils as funded local-compute community hubs and a sustainable public service.
  • Set real KPIs/numbers later โ€” primarily node count and total addressable data, plus event attendance โ€” and fold them into the project plan when the milestone is scheduled.
  • Anchor the launch initiative to November 5th, 2027.

See

Transcript

๐Ÿ“Ž sovereign-compute