Hello, here's the core dev meeting of the month, I apologize we lost 5 minutes early on as there was a recording issue

AI generated Meeting Summary

Because it's AI generated some mistakes may have slipped through, please listen to the meeting for the accurate info

1. HiveMind Sync Time & Robustness

HiveMind full sync time reduced from ~2.5 days to ~24 hours, significantly shortening the change/test cycle.
Multiple robustness improvements added to make HiveMind more resilient against various crash modes.
Outstanding HiveMind issue seen in RC-12 replay (block processing + DB issue) is expected to be resolved by the upgrade.

2. PostgreSQL 18 Status

Codebase is now Postgres 18 compatible, but this release still ships on Postgres 17.
Delay was due to a dependency lacking PG-18 support at the time; it may be ready now, but too late to swap in. Next release expected to ship on 18.

3. K6 API Performance Testing

K6 API performance testing added across all HAF apps.
Not heavily used this cycle, but intended to catch regressions in API calls going forward.
Additional motivation: gives AI measurable data to verify whether its changes improve or regress performance.

4. Release Schedule & Testing Tags

Target: release next week. App owners asked to test against apauditsyncat.com as soon as possible.
Tag naming clarified: RC-12 is old (weeks back). Official 1.x-style tags are protected and can't be easily re-tagged, so trivial changes forced fresh RC numbers, reaching 12.
Host moved to test2 / test3 tags to avoid that churn and the downstream ENV file updates it caused.
Use test2 for latest testing (test3 also acceptable). Real RC pending; HiveSense is the last blocker.

5. Feed Algorithm & Post-Subscription Work

Primary focus for the host: custom feed algorithm feature.
Additional minor work on the post-subscription feature.
Instrumentation added to Hive where it was missing on some APIs.
Ongoing rebasing and progression of open HiveMind features.

6. Boost / Compilation Updates

Boost updates are not for building against Ubuntu 24.
Aimed at downstream ideas around speeding up compilation. Too early to discuss details publicly.

7. Block Explorer - TVL Widget

api.syncad now exposes Total Value Locked on Hive.
New widget on Block Explorer shows total value locked, broken down into HP, HBD savings, and Hive savings, displayed in both Hive units and USD.
Positioned as useful for DeFi watchers.

8. Block Explorer - Other Updates

Updated wax and health-checker versions.
Minor layout issue noted in health-checker: dark background from a denser theme doesn't render correctly inside Block Explorer. Expected easy fix.
Duplicate entries bug (two entries showing for current date) patched on the frontend; backend API issue to be filed separately.
Gandalf's key display request implemented: UI now shows full vs shortened key depending on screen/container size.
Transaction statistics chart added, using HAF's daily volume API. Also pushed to HiveScan. Known rounding issue filed against Block Explorer.
Block Explorer team remains resource-constrained, no funding; progress is incremental.

9. Spear - Claude-Driven Unity Automation

Started with a Codex + MCP-style integration for Unity; switched to Claude (Max), which worked significantly better.
Current setup runs end-to-end automatically: boots Mongo, boots the backend, boots the Unity editor, makes changes, runs the editor, tests on its own.
Limitation: no live visual feedback. Uses screenshots with 5-10s waits, presses buttons via Unity Execute Code (same path as a real click), then re-screenshots to verify state.
Delivered 3-4 features (frontend + backend) in a single day.
Visual polish still needs a human; code wiring does not. Viewed as a directional shift for game development.

10. V2V (Hive <-> Lightning) Rewrite

V2V rewritten from scratch; entirely new backend.
Everything now drivable via custom JSON operations on Hive - not widely known yet.
Waiting on VSC to expose a Hive/HBD-vs-Bitcoin pool. Once available, the current balance-display frontend can be largely retired and custody removed.
In the meantime: balances on the system can be moved, used to pay Lightning invoices, or sent externally, all via custom JSON. Fully documented; contact Brian for integration help.

11. V2V Fee Model Change

Previous model charged a meaningful fee per small sat send.
New model: deposit once (e.g. $100), then fan out 1,000-sat sends via custom JSON, with the fee only applied at deposit.
User-to-user on the system: no fee.
User to external Lightning: usually very small fee; node deliberately run on a low-fee basis.
Rewrite not yet publicised; announcement and writeup pending.

I will Vote for you if you can get me removed from the Automatic 7 Day Downvote Blackist … sponsored by Buildawhale and Blocktrades ?

I have been blogging here since 2017

Everything downvoted since 2021

Hey @offgridlife, I've been a core developer on Hive for six years now, shipping features like recurrent transfers, RC delegations, the Mesh API (Coinbase integration), and most recently several HF28 features. I've also been reworking communities and notifications in Hivemind, and I'm currently building a "for you" feed to solve content discovery. I will also be working on HBD payment rails (subscriptions, a decentralized Patreon/Substack) to make HBD more useful and drive HIVE demand. As my previous proposal is expiring, I've decided to make a new one to work another year for hive while reducing my salary from 350 HBD to 300 HBD to support the current belt tightening efforts. Would you mind reviewing my new proposal and consider voting for it if you find it worthy of your support ? https://peakd.com/proposals/371 Thanks a lot!

Sort:

Trending

[-]

asgarth (69) 18 hours ago

Thanks for the recap. Do you have a gitlab link for the post-subscription feature?

$0.12

2 votes