GrahamBell

Time = Power. A new Sybil-Resistance blockchain model where influence is earned through time not bought.

Try Murphy — free calls that pay you. 🕧

[GrahamBell @GbellOfficial]

Join our discord:

https://t.co/9bUnh0rplm

[GrahamBell @GbellOfficial]

5/ Thresholds were locked before the runs, and every table reproduces byte for byte.

Full write-up: https://t.co/0XcvuJPj43

Code: https://t.co/fD7nOh0VPN

Claim Your ID: https://t.co/mVrSBjX0Xu

#Blockchain

[GrahamBell @GbellOfficial]

Milestone 3 results: does GrahamBell survive the real world?

100,000 simulated miners, real latency between 8 regions, packet loss, clock errors and correlated household outages. A thread on what held, what broke and what changed.

[GrahamBell @GbellOfficial]

4/ Limits: a ~11 h blackout of a region holding ~38% of witnesses pauses issuance a median ~7.6 min (longest 49), then it recovers. Correlated outages leave ~6% of witness groups below quorum. Distant regions win up to ~3% less.

[GrahamBell @GbellOfficial]

Milestone 3: 100,000 simulated miners, real network delays, household outages, and what held, what broke, and what we changed!

[Hurairah Shamsi @HurairahShamsi_]

GrahamBell Milestone 3a and 3a.1

Does it survive the real world?

GrahamBell caps every miner at 1 hash attempt per second and issues one identity every 30 seconds network-wide. Parts 1 and 2 tested the maths and the cryptography in an ideal world. Milestone 3 removes the ideal world: 100,000 simulated miners, real latency between 8 regions, packet loss, clock errors and correlated household outages.

What held:
• The simulator reproduces the Part 1 maths: 51% of active IDs at block 11,362 vs 11,336 predicted
• Blocks every 30.23 s under realistic churn (target 30 s)
• Withholding blocks earns about 15% of a fair share
• Restarting for lucky starts earns under a tenth
• Difficulty hopping gains at most 1.7%, inside the pre-locked 2% limit
• Honest bans for being offline: 0%

What broke, and was replaced:
• A penalty rule that would have banned honest households for power cuts
• A timestamp tie-break that caused deep reorgs (now first-seen, as in Bitcoin)
• A start rule that let late miners catch up instantly
• Permanent rejection of early blocks, which split validators under clock errors

The simulator also found 5 bugs in itself, including one that made our results look better than they were.

Limits, stated openly:
• A ~11 h blackout of a region holding ~38% of witnesses pauses issuance a median ~7.6 min (longest 49 min), then it recovers
• Correlated outages leave ~6% of witness groups below quorum, vs <1% under independence
• Distant regions win up to ~3% less

Thresholds locked before the runs. Every table reproduces byte for byte. Open source.

Full write-up: https://t.co/T44zO9FKZu

Code: https://t.co/gRSKYN9njs

Claim Your ID: https://t.co/dNbfFNJZAZ

Next: Milestone 3b, the attackers. Try to break it.

#Blockchain #Crypto

[GrahamBell @GbellOfficial]

Our founder's take on capital leaving crypto. We agree: it's a chance to fix what was never fixed. Decentralisation, security, and power that money can't simply buy. GrahamBell is building the foundation first, tested in public, failures published too.

[Hurairah Shamsi @HurairahShamsi_]

Honestly, I see this as GOOD news for blockchain.

I said back in 2020 that if this industry continued down the path it was taking, the technology itself would eventually pay the price. We spent years financialising the promise of blockchain without solving some of the fundamental problems that have existed since day one.

Endless token launches. Incentives disguised as adoption. Airdrops mistaken for product market fit. Valuations mistaken for technological progress.

And now people are surprised that the easy money is drying up?

Good. Let it.

If the departure of speculative capital forces this industry to confront reality, blockchain might finally have a chance to fulfil its original promise.

Can we finally stop asking how to manufacture the next token narrative and start answering the questions that actually matter?

How do we achieve genuine decentralisation, robust security, AND scalable on-chain execution without sacrificing one to obtain the others?

How do we prevent capital, specialised hardware, and concentrated infrastructure from turning supposedly decentralised networks into systems controlled by whoever can afford the most resources?

These are not new problems. They have been there since day one. Yet too much of the industry has spent its energy optimising token economics, chasing liquidity, and selling the next cycle instead of confronting them.

This is precisely the kind of problem we’re working on at GrahamBell.

We're starting with the foundation and leaving scalability until the core thesis has survived testing.

We’re researching whether time and sustained infrastructure participation can provide a different foundation for Sybil resistance and influence in a blockchain network, challenging assumptions about the relationship between hardware, capital, and power.

The research is ongoing. The simulations and adversarial testing matter more than the pitch, and the results must stand up to scrutiny.

I’m not interested in building another token narrative to ride the next bull market.

I’m interested in solving problems that blockchain has struggled with since its inception.

Let speculative capital leave. Let weak ideas die. Let projects that exist only to extract value disappear.

If blockchain technology is worth saving, it’s time we started building systems worthy of the word decentralisation.

https://t.co/gRSKYN9njs

[GrahamBell @GbellOfficial]

Can an attacker cheat the clock?

[Hurairah Shamsi @HurairahShamsi_]

Can an attacker cheat the clock?

GrahamBell caps every miner at 1 hash attempt per second. Last week's math assumed that cap can't be gamed. Milestone 2 tested it with real cryptography: real SHA-256 chains, real BLS signatures.

How it works: each round, a miner's header is locked with randomness made from its witnesses' signatures. Every attempt is chained to the last and tied to the next second. You can compute your whole chain instantly, but that only tells you WHEN you'll win. Witnesses sign only when their clocks reach that second, and validators reject blocks that arrive early.

So: can you re-roll the randomness instead?

→ Earlier versions of my design: yes. 4× faster with 4 tries, 15× with 16, 41× with 64.
→ Current design: 1.01× (95% CI 0.93–1.09). BLS signatures are unique, so re-signing gives the same randomness every time.

Also measured:
- Double headers in one round: 500/500 caught, zero false positives
- Old sort-before-hashing rule: 36 headers per hash → replaced with fixed field order
- Withholding a block to re-roll seat placement: ×1.10 at 10% of blocks, and it costs the attacker the block

The weakness we found, said openly:
A witness colluding with a miner can choose to sign or stay silent, giving the miner two randomness options: 1.88× per colluding witness. The fix is threshold BLS (any two-thirds produce the identical signature). Prototype: ~0.5 s setup per member at 40 members. Next stage tests it.

Locked before simulation:
- Pass/fail thresholds for all 9 hypotheses, dated
- Fork-choice loophole closed
- Witness double-signing now bannable

277 tests, 35 cross-checks, byte-identical on a second machine. Open source.

Full write-up: https://t.co/3RwzuP6SPx

Code: https://t.co/gRSKYN9njs

Next: Milestone 3, the issuance simulator. Try to break it.

[GrahamBell @GbellOfficial]

https://t.co/fD7nOh0VPN

[GrahamBell @GbellOfficial]

Full write-up: Can time, instead of money, secure a blockchain? First results from GrahamBell https://t.co/MO5Y1wai2I

[Hurairah Shamsi @HurairahShamsi_]

Can time, instead of money, secure a blockchain?

I'm building GrahamBell, a Layer 1 where new identities are issued at a fixed global rate: one every 30 s, ~1.05M a year, however much hardware shows up. Each participant gets 1 hash/second, enforced by groups of other participants.

The idea (Proof of Infrastructure Endurance / Proof of Time): influence can't be bought in a burst. It has to be built up over years.

That's a claim. So I'm testing it in public. First milestone results, exact math, not yet a network simulation:

→ An attacker winning share s of new IDs approaches a ceiling and never passes it. With all honest IDs online, that ceiling is s. Win 30% of new IDs, end at 30%. Never 51%.

→ An attacker winning EVERY new ID still needs 2.87 years to reach 51%, from 2.9M genesis IDs.

→ Honest uptime matters: with only half of honest IDs online, a 40% attacker passes 51% after 16.4 years.

The math also broke parts of my design. Fixed:
- Restart attack: 150–182× advantage → exactly 1×
- Committee capture at 25%: >99.9% → ~2 in a million
- Bitcoin-style difficulty hopping: +100% → removed
- Seat reshuffling on downtime: cut to 3.6%

Still weak, and said openly:
- An attacker with 25% can stall ~10% of witness groups by refusing to sign. Local (miners move in ~450 s), and it assumes refusing is free. The protocol bans refusers; next stage tests whether that's enough.
- The current quorum needs honest witnesses online 90.75% of the time. A simpler rule needs 81.5%. We'll test it.

Big caveat: all of this assumes the 1 hash/s rate limit holds. Testing that is next.

130 independent cross-checks, reproducible bit for bit, open source:

https://t.co/C2sGIy5sAy

Tear it apart. That's the point.

Full write-up: https://t.co/uehNWsmIjL

Learn more: https://t.co/oMuJHmftqE

[GrahamBell @GbellOfficial]

👀

[Hurairah Shamsi @HurairahShamsi_]

For the past couple of years I've been designing https://t.co/y0zmkgkf2C, a Layer 1 blockchain built on one question: can influence in a decentralized network be tied to time instead of money?

In Proof of Work, capital buys hashpower. In Proof of Stake, capital buys stake. Either way, whoever has the most money can acquire control quickly.

GrahamBell takes a different approach. Every miner is capped at 1 hash per second, so faster hardware gives no advantage. New identities are issued one at a time, about one every 30 seconds across the whole network. An attacker can still run many machines, but each only earns a proportional share, slowly, while honest participants compete alongside it. The claim: under honest competition, taking over the network should take years of sustained effort, not a one-off purchase. I call it Proof of Infrastructure Endurance and Proof of Time (A New Time Based Sybil Resistance Model).

A claim like that means nothing until someone tries to break it. So I'm building an open simulation in Rust that deliberately attacks the design: patient long-term attackers, botnets, colluding witnesses, network failures.

Three commitments:
→ hypotheses and pass/fail thresholds published before the results
→ code public, so anyone can rerun every number
→ results published whether they support the model or not

First results, the exact mathematical baseline before any simulation, are coming soon.

If you think you can break it, I want to hear how.

[Hurairah Shamsi @HurairahShamsi_]

For the past couple of years I've been designing https://t.co/y0zmkgkf2C, a Layer 1 blockchain built on one question: can influence in a decentralized network be tied to time instead of money?

In Proof of Work, capital buys hashpower. In Proof of Stake, capital buys stake. Either way, whoever has the most money can acquire control quickly.

GrahamBell takes a different approach. Every miner is capped at 1 hash per second, so faster hardware gives no advantage. New identities are issued one at a time, about one every 30 seconds across the whole network. An attacker can still run many machines, but each only earns a proportional share, slowly, while honest participants compete alongside it. The claim: under honest competition, taking over the network should take years of sustained effort, not a one-off purchase. I call it Proof of Infrastructure Endurance and Proof of Time (A New Time Based Sybil Resistance Model).

A claim like that means nothing until someone tries to break it. So I'm building an open simulation in Rust that deliberately attacks the design: patient long-term attackers, botnets, colluding witnesses, network failures.

Three commitments:
→ hypotheses and pass/fail thresholds published before the results
→ code public, so anyone can rerun every number
→ results published whether they support the model or not

First results, the exact mathematical baseline before any simulation, are coming soon.

If you think you can break it, I want to hear how.

[GrahamBell @GbellOfficial]

Time = Power!

A New Time-Based Sybil Resistance Model with Capped Proof of Work!

https://t.co/nyHJNw6BLO

[GrahamBell @GbellOfficial]

A different perspective on decentralization. 👀

[Hurairah Shamsi @HurairahShamsi_]

One point from this thread that stands out:

Node count ≠ decentralization.

If one actor ultimately controls the influence, it doesn’t matter whether it’s exercised through 1 node or 10,000 nodes.

Read the thread below 👇

[Hurairah Shamsi @HurairahShamsi_]

The beauty about Time is that it cannot be:

Bought
Accelerated
Compressed
Parallelised

You're stuck in the same sequential, per-second pacing reality, no matter how much money you have.

So tying Time (endurance) with Security makes sense. What do you think?

[Hurairah Shamsi @HurairahShamsi_]

In Proof of Work (PoW): More hardware and electricity = Increased Power / Influence

In Proof of Stake (PoS): Scaled Capital = Increased Power / Influence

Money can buy both almost instantly!

But what if Power was attached with TIME itself?

[Hurairah Shamsi @HurairahShamsi_]

Every blockchain asks:

“How much hardware do you have?”

or

“How much money do you have?”

GrahamBell asks:

“How long have you stayed?”

[Hurairah Shamsi @HurairahShamsi_]

Most blockchains reward accumulation.

GrahamBell (@GbellOfficial) rewards sustained infrastructure endurance and participation over TIME.

[Hurairah Shamsi @HurairahShamsi_]

I've spent 5 years on a question most people don't ask:

Can decentralization survive without rewarding accumulation?