I wrote this after building two fairly large products with Cursor over the last six months. Each ended up with ~40 database tables, payments, email infrastructure, and other typical backend pieces.
Most digital memory today is stored as timelines - feeds, profiles, chronological posts. We’re experimenting with a spatial model instead.
Honoramma lets people create isometric “memory parks”: grid-based maps where monuments can be placed in space. Parks can be private (family-only) or public. Monument slots are activated with a one-time purchase rather than a subscription, because we didn’t want memory to depend on recurring payments.
Under the hood:
- Next.js frontend
- Supabase (Postgres + pgvector)
- Custom isometric renderer (Pixi.js)
- Structured export format for long-term preservation
- Internal ledger system for slot activation
The core question we’re exploring:
Are spatial systems more durable for long-term digital memory than feed-based ones?
I’d really appreciate feedback on the architectural approach and long-term durability of platforms like this.
Happy to answer technical or conceptual questions.
Interesting approach. How do you handle long-term drift or incorrect pattern capture?
If the system mines behavioral patterns from decisions, I imagine there’s a risk of reinforcing mistakes over time. Do you have a mechanism for pruning, versioning, or validating learned memory before it propagates across agents?
When all 21M BTC are mined (around 2140), block rewards go to zero and miners are compensated only through transaction fees.
The network security model then relies entirely on fee incentives. If transaction volume remains high enough, miners stay profitable. If not, hash power could decrease, which might affect security assumptions.
As for hardware: most modern BTC mining uses ASICs, which are highly specialized and not practical for AI workloads. They’re optimized for SHA-256 hashing and can’t be easily repurposed. GPUs, on the other hand, can be redirected to AI or other compute-heavy tasks.
The bigger open question is whether the long-term fee market will be sufficient to sustain miner incentives without block subsidies.