diff --git a/handbook/troggle/trogspeculate.html b/handbook/troggle/trogspeculate.html
index 30fa2e14d..f45413cbd 100644
--- a/handbook/troggle/trogspeculate.html
+++ b/handbook/troggle/trogspeculate.html
@@ -131,6 +131,7 @@ could mean that we could gain.
 But we could move to a read-only system where the only writing happens on data-import.
 Then we could use python 'pickle()' or 'json()' read-only data structures, but we 
   would need to create all our own indexing and cross-referencing code (which is <a href="#mud">a much bigger job</a><sup>*</sup> than you might think - in fact, it is impossible in a standard wsgi/asgi architecture where there is no shared memory between users and each session is independent).
+  <p>[However, in 2026 an AI can cope with this. This has not gone unnoticed: <a href="https://mkennedy.codes/posts/raw-dc-the-orm-pattern-of-2026/">Raw+DC: The ORM pattern of 2026?</a>.]
 <p>There would be more lower-level code, but the
 different segments of the system could be in caving-sensible modules not
 django-meaningful modules. And we would not have all the extra