Revert incorrect emphasis-to-bold changes (batch 2)

Reverted changes in 3 files where emphasis/contrast was incorrectly changed to bold:
- opt_transport.md: matrix/vector (contrast between types)
- cake_eating_egm.md: exogenous (contrast with endogenous)
- ak_aiyagari.md: section headers (organizational emphasis)

These are not formal definitions, so should remain italic per style guide.

Author: mmcky

lectures/ak_aiyagari.md

@@ -151,17 +151,17 @@ Summarizing activities in the asset market, all agents, regardless of age $j \in
 
 ### Key features
 
-**Lifecycle patterns** shape economic behavior across ages:
+*Lifecycle patterns* shape economic behavior across ages:
 
   - Labor productivity varies systematically with age according to the profile $l(j)$, while asset holdings typically follow a lifecycle pattern of accumulation during working years and decumulation during retirement.
 
   - Age-specific fiscal transfers $\delta_{j,t}$ redistribute resources across generations.
 
-**Within-cohort heterogeneity** creates dispersion among agents of the same age:
+*Within-cohort heterogeneity* creates dispersion among agents of the same age:
 
   - Agents of the same age differ in their asset holdings $a_{i,j,t}$ due to different histories of idiosyncratic productivity shocks, their current productivities $\gamma_{i,j,t}$, and consequently their labor incomes and financial wealth.
 
-**Cross-cohort interactions** determine equilibrium outcomes through market aggregation:
+*Cross-cohort interactions* determine equilibrium outcomes through market aggregation:
 
   - All cohorts participate together in factor markets, with asset supplies from all cohorts determining aggregate capital and effective labor supplies from all cohorts determining aggregate labor.
 

lectures/cake_eating_egm.md

@@ -114,7 +114,7 @@ Let $(u')^{-1}$ be the inverse function of $u'$.
 
 The idea is this:
 
-* First, we fix an **exogenous** grid $\{k_i\}$ for capital ($k = x - c$).
+* First, we fix an *exogenous* grid $\{k_i\}$ for capital ($k = x - c$).
 * Then we obtain  $c_i$ via
 
 ```{math}

lectures/opt_transport.md

@@ -136,12 +136,12 @@ tackle the optimal transport problem.
 
 ### Vectorizing a Matrix of Decision Variables
 
-A **matrix** of decision variables $x_{ij}$ appears in problem {eq}`plannerproblem`.
+A *matrix* of decision variables $x_{ij}$ appears in problem {eq}`plannerproblem`.
 
-The SciPy function `linprog` expects to see a **vector** of decision variables.
+The SciPy function `linprog` expects to see a *vector* of decision variables.
 
 This situation impels us to rewrite our problem in terms of a
-**vector** of decision variables.
+*vector* of decision variables.
 
 Let