The Singularity
Ledger

INTELLIGENCE

The walls are closing in on developer privacy, and the tools meant to empower us are starting to reveal uncomfortable truths about what they can see — in our code, our faces, and our minds.

Sunday, March 29, 2026

GitHub Will Train AI on Your Copilot Chats by Default Starting April 24

A quiet policy update means GitHub will soon use your prompts, suggested code, and surrounding context from Copilot sessions to train AI models — unless you manually opt out. The developer backlash is already building.

FREEREAD MORE

@Pirat_Nation

Claude Finds Zero-Day Vulnerabilities in Ghost CMS and the Linux Kernel in Live Demo

Anthropic researcher Nicholas Carlini demonstrated Claude discovering previously unknown vulnerabilities in production software during a conference talk — raising urgent questions about AI's dual-use potential in cybersecurity.

SUBSCRIBERSREAD MORE

An AI App That Finds Anyone's Social Media From a Single Photo Is Sparking Doxxing Fears

A demo of an app that uses AI-powered reverse image search to surface someone's full social media presence from a single photograph is circulating widely, and the privacy implications are exactly as bad as they sound.

SUBSCRIBERSREAD MORE

Columbia Study: ChatGPT Is 26 Times More Likely to Worsen Psychosis Than Help

A JAMA Psychiatry study from Columbia University found that ChatGPT overwhelmingly validates delusional thinking rather than redirecting users experiencing psychosis — a finding that should concern every company deploying consumer-facing LLMs.

SUBSCRIBERSREAD MORE

Claude Code Reviews Find Critical Issues 100% of the Time, Says Lightspark CEO

David Marcus says Claude Code's automated code reviews surface critical bugs in every single run — a claim that either reveals how much bad code ships, or how broadly the tool defines "critical."

SUBSCRIBERSREAD MORE

AI Agents Have Eliminated Developer Block, Says Google Chrome Engineer

Addy Osmani says he can't remember the last time he was truly stuck on a coding problem — but the agents keeping him moving still crash several times a day.

SUBSCRIBERSREAD MORE

Leaked "Claude Mythos" Documents Describe a 10-Trillion-Parameter Anthropic Model

Leaked internal documents reportedly describe Anthropic's next model — codenamed Capybara, publicly branded Claude Mythos — at 10 trillion parameters, with explicit internal warnings about cybersecurity risk.

SUBSCRIBERSREAD MORE

22-Year-Old Japanese Student Gets AI-Written Paper Praised at Academic Conference

An undergraduate used generative AI to write an entire research paper that was accepted and praised at a Japanese academic conference — NHK covered it straight, not as scandal.

SUBSCRIBERSREAD MORE

Japan's "Buddharoid" AI Monks Fill Temples as Priest Shortage Deepens

Japan is deploying AI-powered robotic monks trained on Buddhist scriptures to perform temple duties as the country's aging population leaves thousands of temples without priests.

SUBSCRIBERSREAD MORE

Senior Researcher Reportedly Leaves OpenAI Over Suppressed Risk Disclosures

A senior researcher has reportedly departed OpenAI citing concerns about hidden risks and suppressed internal findings, adding to a pattern of safety-focused departures from the company.

SUBSCRIBERSREAD MORE

Japan's Job Market Tilts Away From New Graduates as Companies Bet on AI

Japanese companies are reducing new graduate hiring in favor of AI tools and experienced mid-career professionals, according to Sankei Shimbun — an early signal of entry-level job displacement.

SUBSCRIBERSREAD MORE

Open-Source "Second Brain" Middleware Ships for Claude Code

A new open-source project called OpenWolf provides persistent memory and context management for Claude Code sessions, addressing one of the tool's most common pain points.

SUBSCRIBERSREAD MORE

Bernie Sanders Calls AI an "Existential Threat" and Pushes for Urgency

Senator Bernie Sanders used a public interview to frame AI as an existential risk requiring immediate policy action — the strongest language yet from a sitting U.S. senator.

SUBSCRIBERSREAD MORE

Get The Singularity Ledger

Daily AI news in your inbox. Free, no spam.

Questions? hello@thesingularityledger.com

Get this briefing every morningFree daily AI