<p><span class="h-card" translate="no"><a href="https://mastodon.social/@neuronakaya" class="u-url mention">@<span>neuronakaya</span></a></span><br />get well soon</p>

<p>I got sick for two nights and was trembling from the cold. I had terrible headaches that made me cry for hours. I hate how my brain brought up depressing memories</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@<span>whitequark</span></a></span> well there's been <code>auto</code> and similar tactics in stuff like Rocq for forever, it's not like people haven't tried to automate it</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> </p><p>c.peeDaig ?</p><p>Maybe the LLM is reading in reverse?</p>

<p><span class="h-card" translate="no"><a href="https://inherently.digital/@hierarchon" class="u-url mention">@<span>hierarchon</span></a></span> that approach is sound although using LLMs for that is stupid; if you cross out &quot;LLM&quot; in it then this is just using an SMT solver</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@<span>whitequark</span></a></span> i know I've seen terence tao throw around the idea of asking LLMs to generate machine proofs in Lean or whatever, since doing the actual formalization is often annoying and the result is verifiable</p>

<p><span class="h-card" translate="no"><a href="https://meow.social/@mimir" class="u-url mention">@<span>mimir</span></a></span> it&#39;s creating and calling a script for its own shell</p>

<p>this is the first or maybe fifth step on a journey that eventually leads to &quot;just feed the netlist into an LLM and ask it to make it smaller, then use SAT to check if it&#39;s a legal transformation or not&quot; (an actual thing i&#39;ve read a paper on recently)</p>

<p>Still can&#39;t get over how 5 years ago the natural history museum of the University of Zürich, an otherwise outstanding place of scholarship, displayed an UNICORN.</p>