Link ar1_turningpts to ar1_bayes and improve exposition

[jstac]

What

Better links ar1_turningpts (Forecasting an AR(1) Process) with ar1_bayes (Posterior Distributions for AR(1) Parameters), reduces duplicated exposition, and tightens the lecture's beginning and (previously abrupt) ending.

Changes

Cross-linking

• ar1_turningpts now references ar1_bayes in the overview, model setup, and posterior section, and treats the posterior construction more briefly rather than re-deriving it.
• Frames the posterior computation as the conditioning assumption of ar1_bayes — apt here, since the initial path starts from an atypical y_0 = 10.
• Added a forward {seealso} from ar1_bayes back to ar1_turningpts.

Exposition

• Matched the AR(1) model introduction to the bulleted style used in ar1_bayes.
• Added a Conclusion and interpretive text after the original/extended Wecker figures, tying results back to the two sources of uncertainty promised in the overview (fixes the abrupt ending).
• Defined the negative turning point statistic N_t, which was referenced but never defined.

Style / cleanup

• Converted the standalone initial-path figure to a MyST mystnb caption (dropped the embedded ax.set_title); kept per-subplot titles on the multi-panel grids.
• Fixed typos: i.i.d. → IID, "we apply we apply", "intergration".

🤖 Generated with Claude Code

[jstac]

Superseded by #969, which redoes these edits on top of the numpyro/JAX rewrite of the lecture (#584). This branch was based on the old PyMC version and had become unmergeable.