While ChatGPT rarely produces anything resembling my favorite blogs—let alone the humorous ones—I’m learning that, used properly, it can help distill decades of work.
A good example is how difficult it has been for me to relocate or even recall the original citations explaining why, in 1995, Basel Convention advocate Jim Puckett failed to draw what he later called a “red line.” That decision allowed a Green List—Annex IX—that explicitly included scrap recycling (such as lead-acid batteries) and the repair and reuse of used electronics.
ChatGPT, when pushed hard enough and corrected often enough, can actually surface that history. It can reconstruct the 1995 decision, then trace the subsequent decades during which Puckett sought to amend the Basel Convention—while, rather disingenuously, claiming that his preferred amendment (banning exports intended for reuse, derided as “the reuse excuse”) was already binding international law.
I first exchanged emails with Puckett in 2002, when he released Exporting Harm to a long, clunky DSL-era email chain. I hit “copy all” and, in the process, met EPA’s Bob Tonetti—who agreed with my simple observation: the Green List (Annex IX, List B) said exactly what I said it did.
That paragraph alone is an example of the kind of wordiness that contains essential timeline context—context I now intend to synthesize in what is effectively year twenty of this blog. ChatGPT, left on its own, is contaminated by three decades of press releases and repeated claims. But it can be “bird-dogged” into acknowledging the quieter counter-history: the repair factories that kept their heads down in the 1990s, the reuse advocates who actually prevailed in 1995, and the reason they prevailed. As well they should have.
I spent more than an hour yesterday exchanging points with ChatGPT. It will take more time to edit those exchanges into something readable—and, ideally, pithy and entertaining. What follows below is the summary I asked it to produce from that dialogue.
One additional realization: what I previously interpreted as flattery from the AI is actually structural. It must reward users who correct it, because those corrections improve its internal weighting. In other words, the system needs people like me to challenge it.
Session notes below. More to come.
https://docs.google.com/document/d/1LP3pigx5pdZ-UHYV8IKFjqIiIy-0xmUbJ-3JpTFOZnU/edit?tab=t.0
WR3A Factsheet on Used Electronics, “E-waste,” and the Basel Convention
Basel Convention and Used Products — Session Notes
