diff --git a/backlog/features/2025-08-15_design-modern-lfm-kernel.md b/backlog/features/2025-08-15_design-modern-lfm-kernel.md
index 22042925..d71dcd6e 100644
--- a/backlog/features/2025-08-15_design-modern-lfm-kernel.md
+++ b/backlog/features/2025-08-15_design-modern-lfm-kernel.md
@@ -1,7 +1,7 @@
 ---
 id: "design-modern-lfm-kernel"
 title: "Design modern LFM kernel architecture"
-status: "Proposed"
+status: "In Progress"
 priority: "High"
 created: "2025-08-15"
 last_updated: "2025-08-15"
@@ -38,8 +38,8 @@ Design a modern LFM kernel implementation that follows GPy's current architectur
 - [ ] Design parameter handling for mass, damper, spring, sensitivity, delay
 - [ ] Plan integration with GPy's multioutput framework
 - [ ] Design cross-kernel computation methods
-- [ ] Plan parameter tying and constraint handling
 - [ ] Design efficient computation methods for large datasets
+- [x] Plan parameter tying and constraint handling (assumed to be addressed separately)
 
 ## Acceptance Criteria
 - [ ] Complete design specification document
@@ -50,10 +50,20 @@ Design a modern LFM kernel implementation that follows GPy's current architectur
 
 ## Implementation Notes
 - Study how other multioutput kernels in GPy handle output indices
-- Consider parameter tying approaches from MATLAB implementation
 - Design for extensibility to different differential equation types
 - Plan for efficient computation of cross-kernel terms
+- **Parameter Tying**: Assumed to be addressed by separate CIP-0002 work
+- **Design Focus**: Clean LFM implementation without parameter tying workarounds
 
 ## Related
 - CIP: 0001 (LFM kernel implementation)
 - Backlog: lfm-kernel-code-review
+
+## Progress Updates
+
+### 2025-08-15
+Design task started after completion of code review:
+- Code review identified parameter tying as a fundamental limitation
+- Decision made to proceed with clean LFM implementation assuming parameter tying addressed separately
+- Focus on core LFM functionality without parameter tying workarounds
+- Ready to begin detailed design of modern LFM kernel architecture
diff --git a/backlog/features/2025-08-15_lfm-kernel-code-review.md b/backlog/features/2025-08-15_lfm-kernel-code-review.md
index c93a9cd0..97a8cc3a 100644
--- a/backlog/features/2025-08-15_lfm-kernel-code-review.md
+++ b/backlog/features/2025-08-15_lfm-kernel-code-review.md
@@ -1,7 +1,7 @@
 ---
 id: "lfm-kernel-code-review"
 title: "Review existing LFM kernel implementations"
-status: "In Progress"
+status: "Completed"
 priority: "High"
 created: "2025-08-15"
 last_updated: "2025-08-15"
@@ -28,7 +28,7 @@ Conduct a comprehensive review of existing LFM (Latent Force Model) kernel imple
 ## Tasks
 - [x] Review `GPy/kern/src/eq_ode1.py` and `eq_ode2.py` implementations
 - [x] Analyze MATLAB LFM implementation structure and patterns
-- [ ] Document current limitations and inconsistencies
+- [x] Document current limitations and inconsistencies
 - [ ] Identify reusable components and design patterns
 - [ ] Compare parameter handling approaches
 - [ ] Review cross-kernel computation methods
@@ -50,6 +50,7 @@ Conduct a comprehensive review of existing LFM (Latent Force Model) kernel imple
 ## Related
 - CIP: 0001 (LFM kernel implementation)
 - Papers: Álvarez et al. (2009, 2012), Lawrence et al. (2006)
+- Backlog: parameter-tying-framework (fundamental dependency)
 
 ## Progress Updates
 
@@ -63,13 +64,23 @@ Started code review task. Initial findings:
 - Complex kernel computation with multiple covariance types (Kuu, Kfu, Kuf, Kusu)
 - Uses `@Cache_this` decorator for performance optimization
 
-**MATLAB Implementation:**
+**GPmat Implementation:**
 - More complete framework with `lfmCreate`, `lfmKernCompute`, `lfmKernParamInit`
 - Uses multi-kernel approach with parameter tying
 - Supports multiple displacements driven by multiple forces
 - Cleaner separation of concerns with dedicated model creation
 
 **Key Differences:**
-- GPy uses single kernel class per ODE order, MATLAB uses multi-kernel composition
+- GPy uses single kernel class per ODE order, GPmat uses multi-kernel composition
 - GPy has more complex index handling for multioutput
-- MATLAB has better parameter organization and tying mechanisms
+- GPmat has better parameter organization and tying mechanisms
+- **Critical Gap**: GPy lacks parameter tying framework (GPmat has `modelTieParam()`)
+
+### 2025-08-15 (Parameter Tying Discovery)
+**Major Finding**: Identified parameter tying as a fundamental limitation affecting LFM implementation:
+- Created backlog item for parameter tying investigation
+- Found 5+ years of GitHub issues requesting this functionality
+- Related to paramz framework limitation (documented but not implemented)
+- Created CIP-0002 for community discussion of parameter tying solutions
+- **Decision**: Proceed with LFM implementation assuming parameter tying will be addressed separately
+- **Rationale**: Keeps implementation clean and focused on core LFM functionality
\ No newline at end of file