Whev can't believe it's been 80 montly meetings already, feels like yesterday when we started those.
AI generated Summary:
Personal update (howo)
Posted MRs for API limitation (memory issue, low priority since drone-protected) and a fix for the recurring-transfer bug.
Rebasing Hivemind MRs.
Side project: a second HiveD implementation in go — replays up to 30M blocks before failing. Curious whether a fresh AI pass with full repo context could outperform the original. Plans to maybe run a top-100 witness on it for the experiment.
Building a HAF app to better support HBD recurring transfers and payment — exposes an endpoint to query whether a given invoice ID is being paid, payment failure counts, etc. (closes a UX gap where devs currently have to walk virtual ops).
Drafting a design doc for encrypted on-chain content (Patreon/Substack-style paid content paid in HBD).
Built a per community notification filter on hivemind
Hive core / release (blocktrades)
- Tagging the release today (RC14 is the final candidate). Pre-replayed snapshot will be uploaded; replay times are much better now.
- Stack split across two servers; monitoring setup is the last piece and is mostly done.
- API node upgrades to coordinate, but stagger them — don't have everyone switch at once unless they have backups.
- Big changes in this release:
- Major haf schema rewrite + many haf apps rewritten to match.
- Hivemind near-complete rewrite — almost entirely SQL now, very little Python left.
- Account-balance refactor (maintenance), build-time optimizations, include-file cleanup — held back to adjacent branches until after release.
- Found new deadlock + index-creation bugs in haf.
- Specialized exceptions added to wax (parseable error handling for apps).
- Denser progressing well; will publicize as alternate to hive.blog and beef up its server.
- Python WorkerBee underway, mirroring the TypeScript version.
Haf apps / Block Explorer (mcfarhat)
- New endpoints on haf block explorer: historical proposal votes per proposal, total wallet addresses (daily granularity), proxy-account sorting, proxy-power API fixes, balance tracker fix for negative savings withdrawals.
- Two new projects spun up: haf-stats and haf-fyp.
- FYP: TikTok-style personalized ranking using Hivesense, votes/comments/community subs as weighted signals; open source, hosted on haf API nodes.
- haf-stats: financial summaries, RC footprint per account, network-wide RC utilization, HP distribution buckets, and many more endpoints.
- Repos created on GitLab but he can't push — needs maintainer permission from blocktrades/Gandalf. Action: blocktrades/Gandalf to grant.
- Block-explorer frontend: TVL card, incoming/outgoing RC delegations with timestamps, new per-account balance breakdown (open orders, pending conversions, savings under withdrawal, escrow), Keychain + Hive Signer login, customizable dashboard, tooltip/crash fixes.
- Pending MRs (4): historical proposal votes, total wallet addresses, proxy-account sorting, proxy-API fixes — Michael to review this week.
- howo to share his Hivemind ranking-algorithm MR with mcfarhat for cross-pollination on the FYP algorithm.
Velocity L2
- Very productive month: ~1.5M lines, ~2000 commits across 12 repos.
- Built an inbox/outbox coordination system between 12 parallel Claude (Opus 4.7) sessions after team-agents broke down past 3–4 agents. Hits weekly Opus limit at ~14h of session time.
- Velocity = new L2 with three pillars:
- Block speed: 0.5 ms via RocksDB + heavy compression — Hive's 3s and Magi's 30s aren't viable for fast-paced game reward loops.
- SDKs for Unity, Unreal, Phaser.js, GDScript — covers PS5/Xbox/Switch/mobile/PC so indie devs don't need to learn 4–5 new languages to ship.
- Namespace isolation — each game gets its own fuel allocation and isolated economy, so a bad actor in one project can't drag down others.
- Custodial-by-default onboarding (lift-and-shift for game players); users get keys back when they want to bridge out to Hive/Solana/etc. Two-way bridge (not just one-way).
- Genesis with 3 validators (2-of-3 multisig on Hive + Solana bridge accounts), speaker becomes a regular validator after 6 months. DevNet → Testnet → Mainnet planned; this week + next is heavy testing.
- blocktrades offered passkey-derived wallets (already in Magi testnet) as an alternative to custodial onboarding — speaker still wants custody-on-rails for game UX.
- Side discussion: speaker asked why Hive serves JSON not Borsh — claimed 30–40% space savings. blocktrades/Bartek: custom_binary ops already exist alongside custom_json (rarely used, possibly unknown), and DB-layer compression with dictionaries removes most ASCII overhead anyway. Bartek suggested the
tnauk/tune library for further JSON-footprint reduction; will share link.
Magi L2
- Took over Magi dev leadership; in the last 2 months: 600% PR shipping rate, 600% net code output, 1000% multi-chain code, ~900 commits, 300 merged PRs across 13 repos, ~250k insertions — ~10x on same team/budget.
- Shipped largest mainnet merge to date: BTC/HBD and HBD/Hive pools live, accessible via Altera.
- TSS (threshold signature scheme) custody live on mainnet — Magi is now the 4th protocol with TSS custody and native BTC swaps without wrapping.
- Off-chain gossip replacing flaky RPC; concurrency hardening; TSS lib v3 upgrade.
- Built ZK proofs for the EVM integration — drops validator hardware cost (no full Eth node needed, just a prover). Forked SP1 succinct repo and took ownership; reusable across Hive.
- EVM integration end-to-end on testnet: ETH + USDC deposits/swaps/withdrawals work without wrapping or custodians.
- Six chain integrations done in code: BTC (live), LTC, Dash, Doge, BCH, ETH/USDC (testnet).
- Altera frontend: complete rewrite done. Yozen joined the team (formerly BigD frontend).
- Built passkey-derived wallets, ZK email-to-email HBD transfers (parses email body for transfer instructions).
- New infra: Magi Mongo indexers, Magi NFT contracts, Magi token contracts.
- Incentive pendulum (Oracle auth + economic security + validator fee distribution split with LPs) in testing — final blocker before audits with BlockTrades.
- Partnerships: Dash announced; BuildX privacy-token decision Tuesday (needs $100k liquidity, 50/50 HBD/their token).
- 20 security review sessions, ~1000 scenarios over 2 months → 50 vulns found, 15 security PRs merged, 25 reports.
- Next: launch EVM + Dash + pendulum, then freeze and move to optimization + partnerships.
- howo's feedback: need many more test files / benchmarks / edge-case coverage in the repos. Elbi agreed; planned for the post-feature optimization phase.
===========
That's it for this month, feel free to ask any questions in the comments.
If you like what I do please consider supporting my proposal: https://peakd.com/proposals/371
This sounds very interesting! Code public by chance?
Code not public yet as it's still missing a lot of critical features (right now it's only replaying, can't do P2P) :)
Can we get these uploaded to 3spk or bitchute, or anything but utub?
It's really hard to beat the simplicity and usability that youtube has. I've tried 3speak but it's still missing so much compared to youtube and most importantly stability(I won't know if a video I shared will be there and loading very quickly in 3 months vs with youtube I'm guaranteed to have it there).
So, put it on utub, and somewhere else for us utub refusers?
Bitchute is very stable.
Didn't know about this one.
It was banned in england for not censoring things.
In some countries both are blocked. Even 3speak in some... the only option is to bring the front ends back into the user side. So the only thing they need is to P2P download to watch.