<oembed><type>rich</type><version>1.0</version><title>Cory Doctorow wrote</title><author_name>Cory Doctorow (npub1yx…75qvv)</author_name><author_url>https://yabu.me/npub1yxzkmtuyctjw2pffp6e9uvyrkp29hrqra2tm3xp3z9707z06muxsg75qvv</author_url><provider_name>njump</provider_name><provider_url>https://yabu.me</provider_url><html>nostr:nprofile1qy2hwumn8ghj7un9d3shjtnyd968gmewwp6kyqpq2spdmtqs75xq7mje3am8jq0nr33rkclkyzxm8r22l7w92f0zqsqsdfepsg Further: the cases against models are only tangentially related to memorization. It&#39;s undeniable that these suits would exist even if models had &#34;guardrails&#34; that *perfectly* prevented them from reproducing their training data verbatim. Memorization is *not* the crux of any of these complaints - *training* is.</html></oembed>