XI'AN · HYPERSCALE FOR RADIX

Hyperscale Weekly

Week #8 · 29 Jun – 5 Jul 2026 · Milestone 1 · Adaptive Sharding · Week 10 · ~month 2 of 4
Split the work, not the world
MILESTONE 1 PROGRESS59%

Previously: flightofthefox had just folded the simulator and production into one shared test-scenario framework, with a clean reshape orchestrator driving both runners.

Building on last week's test consolidation, flightofthefox turned to safety and uncovered the biggest bug yet — a rare way the beacon could finalise the same epoch two different ways across a network split. He reproduced it in the simulator, closed it with a new pool-ratified commit path, and proved the fix in five new mathematical models. A landmark week for the beacon's trustworthiness.

82
-15% vs last wk
commits
+11k 6.1k
+16% vs last wk
lines changed
166
files touched
34+
day commit streak
Commits / day
M
T
W
T
F
S
S
peak 17/day · 7/7 days active
Momentum
Lines changed / week
last 8 weeks
82% Rust7% Config7% Docs5% Specs
What flightofthefox built

Beacon double-commit, found and fixed

flightofthefox reproduced a rare safety bug where the beacon could finalise the same epoch two different ways across a network split, then built a pool-ratified commit path that mutually excludes the old route — closing the door for good.

→ see “The Census” on hyperscale.rs

Straddlers and reshapes, more strongly fenced

Transactions caught in a splitting shard can no longer be evicted while a gate still holds them, and the handoff between old and new committees now waits until the successor is actually live.

→ see “The Will” on hyperscale.rs

Five new proofs in math

Five new formal models now prove core safety rules in math: straddler settlement, cross-shard atomic commit, shard consensus, the weighted-time clock, and the dual-commit fix itself.

→ see “The Proof” on hyperscale.rs
The bottom line
Standout change
The beacon's rare double-commit safety gap is now closed, mutually excluded, and proven in math.
What's next
Likely next: the lottery-style validator shuffling that's the other big piece of M1 still to land, now with a much firmer beacon underneath it.
This week's concept
The CensusThe beacon that tracks validators & topology
The central event was the beacon's double-commit safety gap being reproduced, fixed with pool ratification, and mathematically proven. That's a major step forward for the chain that tracks validators and topology.
hyperscale.rs/census
Jargon, decoded
Pool ratificationA way for validator pools to vote on and approve a beacon block before it counts as final, replacing a less safe older path.
Dual commitA safety bug where two conflicting versions of the same block could be finalised at the same time across a split network.
Where the work landed
dual-commit gap closed with pool ratification
simulation · The Crash Lab
reproduced the partition bug end-to-end
new ratification and straddler data types
Reference — concept map, roadmap & links
The full picture
The System
The whole sharded design at a glance.
The Journey
One transaction's trip through the network.
The Clock
Consensus-attested time across independent shards.
The Census
The leaderless beacon of validators, stake and shards.
The Overlap
Why two conflicting blocks can never both commit.
The Generals
All-or-nothing cross-shard commits, computed not voted.
The Archive
Every published byte is preserved or provably expired.
The Library
All state in one merkle tree; a shard is a subtree.
The Will
How in-flight transactions settle when a shard dies.
The Lottery
Random, ever-moving committees; proven cheats are jailed.
The Triage
Under overload, urgent traffic never waits for bulk.
The Governor
Validator entry is re-priced every epoch, like a market.
The Crash Lab
The simulator that replays any failure byte-for-byte.
The Proof
Proving safety with maths, not just tests.
The Asterisks
Every design's trade-offs — including Hyperscale's own.
The road to mainnet
M1
Adaptive Sharding
~4 mo
M2
Radix Engine
~5 mo
M3
Gateway & API
~3 mo
M4
Validator GUI
~3 mo
M5
Migration
~3 mo
M6
Live support
12 mo
Now: Milestone 1 · Adaptive Sharding — Week 10 · ~month 2 of 4 · 59%. Since day one: 2,324 commits — one developer, fully in public.
Links
Written by AI from public commits, the community chat and hyperscale.rs — may contain mistakes. · generated 2026-07-09