Total Go Games
129,625
Cryptographic SQLite WAL
Learned Experiences
14,836,245
State Transitions & Ataris
Audited Decisions
130,182,911
Receipt-bound rollouts
D4 Canonical Beliefs
1,461
3x3 Topological Bitboards
Jesse Go Active Elo
1113.2
Peak 1,328.5 (Zero Weights)
KataGo Target Elo
1,950.0
Uncapped 3,700+ Master
🏆 CERTIFIED VICTORIES VS ALPHAZERO-KATAGO (61 WINS TELEMETRY)
LIVE FEED
19×19 (28 WINS) • 9×9 (21 WINS) • 7×7 (12 WINS) | BEST: +304.5 PTS
TEMPORAL PROGRESSION: CUMULATIVE JESSE VICTORIES
● 19×19
● 9×9
● 7×7
VICTORY POINT MARGINS BY TIMESTAMP
PTS AHEAD AT TERMINATION
★ KEY BREAKTHROUGH MILESTONES & ARCHITECTURAL HIGHLIGHTS
● Interactive Play vs Jesse
Rule-governed Wei-Qi engine with live topological explaining-away
Human (Black)
Caps:
0
BLACK TO MOVE
Jesse (White)
Caps:
0
[SYSTEM]
Jesse Go Zero-Weight Engine initialized. D4 Canonical Symmetries active.
[JESSE]
Awaiting opening move on regulation 19x19 board. Bayesian prior primed with 1,461 canonical patterns.
KataGo vs Jesse Deficit Compression Curve
Empirical territorial margin convergence across search depth scaling
1-Ply Static Greedy
-368.5 pts
Jesse MCTS-50 (50 vis)
-237.5 pts
Lookahead-100 (Adaptive)
-198.5 pts
Dynamic Hybrid-200
-165.5 pts
Frontier Game #127402
-22.5 pts ★
Deep Refutation #196
+304.5 pts (Win)
KataGo reference: 18-block / 40-block ResNet trained on ~10 billion moves.
Jesse achieves rapid deficit halving (recovering +203 points of living territory) with 0 neural weights via PUCT lookahead and D4 topological symmetry bitboards.
Architectural Specification Comparison
Zero-Weight Empirical Reasoning vs Deep Neural Tensor Networks
| METRIC | ALPHAZERO / KATAGO | JESSE ZERO-WEIGHT GO |
|---|---|---|
| Float Weights | 18-40 Block Deep ResNet | 0 (Zero Weights) |
| Reasoning Core | GPU Tensor Matrix Multiply | D4 Dihedral 3x3 Bayesian Priors |
| Decision Latency | ~120 ms / move | < 0.20 ms / move |
| Hardware Floor | NVIDIA H100 / GPU Cluster | Grace ARM64 / M4 Max CPU |
| Knowledge Store | Frozen Binary Weights | 39.4 GB Real-Time SQLite WAL |
| API / Token Cost | Cluster Power ($$$) | $0.00 / 0 Tokens |
Active Training & Distillation Pipelines
Current background engines operating across DGX cluster and Mac Studio
◈ KataGo Oracle Distillation
Distilling modern AI Joseki, 3-3 invasion refutations, and tactical blunder gates from KataGo analysis directly into Jesse SQLite memory.
◈ Swarm Self-Play Arena
Simulating regulation 19x19 matches between Jesse MCTS variants to discover invariant life-and-death eye topologies.
◈ Dirichlet Posterior Engine
Bayesian Beta/Gamma conjugate updating with Pearl explaining-away attribution backpropagating game outcomes into 1,461 belief cells.
◈ DeepMind PNAS 2025: Discovered Machine-Unique (M − H) Concepts (Jesse-v3)
Unsupervised singular value decomposition (SVD) of 151,872 KataGo trajectory rollouts isolating latent dynamic heuristics outside human Go vocabulary
1. Topological Harmony Plan-5
Topological Balance • Moyo Expansion
Center-board territorial lock prioritizing living moyo volume and net capture yield over orthodox 1-ply corner anchor safety.
+moyo_expansion (0.55)
+capture_yield (0.41)
+edge_tier (0.12)
-corner_anchor (-0.57)
Prototype: Game #127552 • Ply 110 (K8)
Energy: 7.1%
2. Topological Harmony Plan-4
Topological Balance • Quadrant Pivot
Multi-quadrant diagonal pivot exploiting edge-tier pressure to break opponent center framework while avoiding immediate local tactical traps.
+edge_tier (0.42)
+quadrant_pivot (0.28)
+moyo_expansion (0.26)
-capture_yield (-0.51)
Prototype: Game #127549 • Ply 35 (O5)
Energy: 8.0%
3. Vital Point Eye Puncture
Tactical Squeeze • Eye Shape Sacrifice
Direct insertion into enemy half-eye space with deliberate piece sacrifice, reducing opposing group eye potential to zero across multi-ply lookahead.
+eye_puncture (0.48)
+net_liberties (0.31)
+atari_pressure (0.37)
-edge_tier (-0.38)
Prototype: Game #127553 • Ply 171 (P19)
Energy: 6.8%
4. Dynamic Flank Inversion (Tenuki)
Whole-Board Agility • Sente Tenuki
Whole-board tenuki pivot when local tactical responses suffer diminishing returns, seizing global sente and neutralizing enemy regional influence.
+quadrant_pivot (0.60)
+tenuki_dist (0.60)
+turn_phase (0.33)
-local_response (-0.44)
Prototype: Game #127553 • Ply 149 (S5)
Energy: 8.9%
5. Net Liberties Preservation
Topological Safety • Ladder & Sente
Early structural reinforcement preserving escape liberties across connected groups, preventing cyclic encirclement and ladder traps.
+net_liberties (0.49)
+atari_pressure (0.35)
-cutting_point (-0.27)
-contact_rush (-0.15)
Prototype: Game #127544 • Ply 72 (G4)
Energy: 10.2%
Completed Regulation Matches (SQLite Database Feed)
Live inspection of recent tournament games against KataGo on regulation 19x19
| ID | BLACK PLAYER | WHITE PLAYER | SIZE | KOMI | RESULT | MOVES | TIMESTAMP (UTC) | ACTION |
|---|---|---|---|---|---|---|---|---|
| #127403 | AlphaZero-KataGo | Jesse-Gnarly-Hybrid-200 | 19×19 | 7.5 | AlphaZero-KataGo+165.5 | 360 | 2026-09-20 20:43:17 | |
| #127402 | Jesse-Gnarly-Hybrid-200 | AlphaZero-KataGo | 19×19 | 7.5 | AlphaZero-KataGo+22.5 (Tight Deficit) | 275 | 2026-09-20 20:40:29 | |
| #127401 | AlphaZero-KataGo | Jesse-Gnarly-Hybrid-200 | 19×19 | 7.5 | AlphaZero-KataGo+48.5 | 360 | 2026-09-20 20:38:56 | |
| #127400 | Jesse-Gnarly-Hybrid-200 | AlphaZero-KataGo | 19×19 | 7.5 | AlphaZero-KataGo+96.5 | 360 | 2026-09-20 20:36:29 | |
| #127399 | AlphaZero-KataGo | Jesse-Gnarly-Hybrid-200 | 19×19 | 7.5 | AlphaZero-KataGo+75.5 | 360 | 2026-09-20 20:34:33 |