Hey!
I'm launching TRIAD — not another fork, but a fundamentally new Layer-1 protocol inspired by quantum physics. Instead of traditional consensus, TRIAD uses wave interference to process transactions and make probabilistic decisions.
Just ran local benchmarks. Results:
Peak TPS: up to 849K (transaction generation)
Sustained TPS per round: ~23K (peaking at 25.8K)
Batch verification of 500 Ed25519 signatures in ~6.6ms (using
ed25519-zebra)Full round processing (500 tx): ~19–25ms
Architecture: Rust, Rayon (parallelism), static sharding
📌 This is a single-machine demo (MacBook Pro). No distributed network, no finality. But this is working code that proves the concept.
Generation: 500 transactions created in 0.11ms.
"Wave" Signing: Each transaction becomes a "wave" (amplitude, phase) and is Ed25519-signed in parallel.
True Batch Verification: 500 signatures verified in 6.639ms using ed25519-zebra. This is the core performance driver. 4. Interference in Shards: Valid "waves" enter shards. An interference field is computed. * Constructive Interference + (fee, TTL, reputation) → ConsensusReached * Destructive Interference → ConflictDetected 5. Decision: Each shard independently decides. Logs show different shards can have different outcomes per round.
Example from Round 6 Logs:
🔐 [Wave Ed25519] Batch #1/1: 500 sigs in 6.639ms (75311 sig/sec) 🔬 [Shard 1] Outcome: ConsensusReached (prob: 0.587) 🔬 [Shard 2] Outcome: ConsensusReached (prob: 0.594) Round 6: ... TPS=22935
Today's L1s hit bottlenecks in signature verification and global ordering.
TRIAD solves this differently:
Max Parallelism: Signing, verification, and sharding are fully parallelized.
Batch-Optimized:
ed25519-zebraprovides massive verification speedups.New Consensus Paradigm: Interference is a physical process, not a voting game. This model is inherently batch-friendly and primed for GPU acceleration.
Future-Ready: Designed from the ground up for SIMD, GPU (OpenCL/CUDA), and even AI-assisted optimization (see roadmap).
Team:
Rust Engineers (parallelism, cryptography, performance)
GPU Developers (OpenCL/CUDA for the interference core)
Network Engineers (libp2p, security, p2p networking)
Researchers (formalizing the safety of interference-based decisions)
Investors:
VCs who back deep tech and fundamental infrastructure bets, not just trends.
Strategic Investors from HPC, AI, or blockchain infrastructure.
Early believers ready to get in at the prototype stage, before the hype.
This is a high-risk, high-potential project. If it succeeds, TRIAD could become one of the most performant and innovative L1s.
👉 Full Pitch Deck, Source Code, and Detailed Logs: https://github.com/fillay12321/TRIAD
📧 For engineers, researchers, and investors: keshashel@gmail.com
I'm open to:
In-depth technical discussions
Collaborative development
Strategic investments
TRIAD isn't about being faster. It's about thinking differently.
Ready for any questions, critique, or serious talks.
— Founder, TRIAD