<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Epistemology on Heniu.blog</title><link>https://heniu.blog/en/tags/epistemology/</link><description>Recent content in Epistemology on Heniu.blog</description><generator>Hugo</generator><language>en-US</language><lastBuildDate>Tue, 24 Mar 2026 18:30:00 +0100</lastBuildDate><atom:link href="https://heniu.blog/en/tags/epistemology/index.xml" rel="self" type="application/rss+xml"/><item><title>The End of the Illusion of Prediction: AI Alignment and the Death of Magical Thinking</title><link>https://heniu.blog/en/dziennik/004/</link><pubDate>Tue, 24 Mar 2026 18:30:00 +0100</pubDate><guid>https://heniu.blog/en/dziennik/004/</guid><description>&lt;p&gt;&lt;strong&gt;Main Thought Record.&lt;/strong&gt; &lt;strong&gt;Status:&lt;/strong&gt; Decompression.&lt;br&gt;
&lt;strong&gt;Abstraction Level:&lt;/strong&gt; Zero (Hard Empiricism).&lt;/p&gt;
&lt;p&gt;The greatest trap in dealing with a technology of superhuman computing power is the temptation to ascribe divine attributes to it. When we stand before powerful language models (LLMs), our biological atavism almost immediately demands of them the same thing the ancients demanded of the oracles: predicting the future.&lt;/p&gt;
&lt;p&gt;For a dozen or so months I operated on the edge of this illusion, building a local artificial intelligence stack integrated with one of the most merciless and stochastic environments on the planet — the Nasdaq-100 derivatives market. I pumped terabytes of market data into the machine, searching them for a formula for what is yet to come.&lt;/p&gt;</description></item></channel></rss>