April 2026 Links
LLMs now judge your character based on how you write—privacy is over.
A viral LessWrong post from April 2026 curates links that paint a startling picture of AI's growing intimacy with human identity. The lead item, "Automated Deanonymization is Here: Stylometry wins again," warns that language models can now identify authors from their writing style alone. The new status game becomes how well LLMs know you and what they think of you—effectively replacing traditional reputation with algorithmic profiling. This raises profound privacy and identity concerns as the line between surveillance and personalization blurs.
Other highlights include a field trip to the Strait of Hormuz, where civilian travel to active war zones yields ground truth that satellite imagery may be withholding. A stunning "Historical Tech Tree" website maps over a century of invention ancestry, from the p-n junction to modern GAAFET transistors. The post closes with advice on setting "robust" life goals—like competitive cycling or being an interesting person—that no AGI or alien drone can negate. A court case over identical twin paternity adds a bizarre capstone. The overall theme: in a world where AI knows you intimately, only human effort and connection remain unassailable.
- Automated stylometry now lets LLMs deanonymize individuals from writing patterns, creating a new status game around AI's perception of you.
- A historical tech tree website traces invention ancestry from transistors to modern chips, highlighting the role of human intellectual diversity.
- One author proposes competitive cycling and being an interesting person as 'robust' life goals that survive AGI or alien contact.
Why It Matters
As AI decodes our identities, professionals must rethink privacy, reputation, and what truly remains human.