<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: mwarren</title><link>https://news.ycombinator.com/user?id=mwarren</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Fri, 19 Jun 2026 16:50:28 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=mwarren" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by mwarren in "AWS Bedrock to require sharing data with Anthropic for Mythos and future models"]]></title><description><![CDATA[
<p>Zero Data Retention does not mean zero <i>information</i> retention. First, this whole discussion regards AWS Bedrock's straightforward policy for users of OpenAI GPT-5.4 and GPT-5.5 models and Anthropic Claude Fable 5 saying "[inputs and outputs will be retained for up to 30 days](<a href="https://docs.aws.amazon.com/bedrock/latest/userguide/abuse-detection.html" rel="nofollow">https://docs.aws.amazon.com/bedrock/latest/userguide/abuse-d...</a> )."  That's <i>plenty</i> of time for a model training run using those inputs/outputs and, once the <i>information</i> is encoded into model weights, the original training <i>data</i> can be deleted to meet the ZDR contract.
For models without the 30-day retention clause, it's still possible for AWS to route inputs and outputs through a dynamic training system to encode the information into model tensors and then toss out the original "data".<p>Edit: linkify</p>
]]></description><pubDate>Fri, 12 Jun 2026 03:00:18 +0000</pubDate><link>https://news.ycombinator.com/item?id=48499364</link><dc:creator>mwarren</dc:creator><comments>https://news.ycombinator.com/item?id=48499364</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48499364</guid></item></channel></rss>