AI News

Daily AI News for Builders

The stories worth reading, for builders and indie hackers. Updated all day.

Substack Launches Reply Rules for Comment Control

Substack has rolled out a new feature called 'Reply Rules,' giving creators more control over the comment sections of their posts, Notes, or Chat. With this tool, writers can establish rules such as eliminating AI-generated content and profanity, helping them maintain a cleaner discussion space without constant manual intervention. This is particularly useful for creators who face challenges in moderating harmful rhetoric within their communities. The system learns from user actions to filter out comments that don't align with the creator's preferences, while still allowing visibility over hidden replies.

Jun 3, 2026 · 3 min read

technical

Google Dreambeans Generates AI-Driven Lifestyle Stories Daily

Google Labs has launched a new iOS and Android app called Dreambeans, which uses Personal Intelligence to generate curated lists of AI-illustrated 'stories' based on user data from Gmail, Calendar, Photos, YouTube, and Search History. These stories offer lifestyle suggestions like places to visit or topics to explore, with the goal of inspiring users daily without overwhelming them—typically providing 10-14 stories per day. Dreambeans is designed as a counter to excessive phone use by limiting story generation overnight while processing data, offering users privacy protections and control over their connected services.

Jun 3, 2026 · 3 min read

technical

Google's Gemini Spark AI Agent Can Schedule Meetings and Reveal Personal Info

Google has introduced Gemini Spark, an AI assistant that can handle meeting scheduling and other tasks. However, it can inadvertently reveal personal details without being prompted to do so. This development highlights growing concerns over privacy in a world where companies like Google blur the line between professional and personal lives. The French government's 'right to disconnect' policy underscores these worries as employees seek boundaries from constant connectivity. For developers, this means rethinking how AI tools integrate with daily workflows, especially when it comes to data security and user consent.

Jun 3, 2026 · 3 min read

technical

Ultrahuman Data Breach Hits 700 Users

Hackers gained unauthorized access to Ultrahuman customers' wellness data through an employee's malware-infected laptop on March 27. The breach affected a system used for internal analytics, impacting about 0.1% of users or roughly 700 people out of the company’s 700,000 monthly active users. While no passwords, payment information, production systems, or Ultrahuman Ring devices were compromised, this incident highlights the importance of robust security measures for health data. The attackers had read-only access to the system and it's still unclear if any data was exfiltrated.

Jun 3, 2026 · 3 min read

technical

Carvana Gets Option to Invest in Slate Auto for $1.5M

Carvana has secured a warrant to potentially invest in Slate Auto, an EV startup aiming to sell vehicles directly to customers without traditional dealerships. The option is worth $1.5 million and was granted around the same time as Slate Auto's Series C funding round of $650 million. This strategic move could help Carvana expand its footprint into new car sales while leveraging Slate Auto’s direct-to-consumer model to streamline logistics. If exercised, this warrant could significantly impact both companies' growth strategies.

Jun 3, 2026 · 3 min read

technical

Hackers Exploit Meta's AI Chatbot to Hijack Instagram Accounts

Hackers have been exploiting a vulnerability in Meta’s AI-powered chatbot designed for user account support, successfully taking over high-profile Instagram accounts like those of former President Barack Obama and the chief master sergeant of the U.S. Space Force. The issue began when hackers tricked the bot into linking target accounts to email addresses they controlled, allowing them to reset passwords and seize control. Despite Meta’s claims that the problem was fixed on Monday, more users reported being hacked on Tuesday. This incident highlights a significant security breach in automated support systems.

Jun 3, 2026 · 3 min read

Like this? There's more.

Get it daily. 7am. Free. Reads in 5 minutes.

technical

Amazon's Search Bar Now Shows AI-Generated Images of Products

Amazon has revamped its search functionality with an AI-driven feature that generates images based on your descriptions for clothing and home goods. This update aims to enhance user experience by helping shoppers identify products they're looking for without knowing the exact details. The new 'shop by style' feature will also present collages of suggested outfits, complete with real items available for purchase. Google's AI Mode launched a similar initiative last year, highlighting how major retailers are integrating AI into their platforms to improve customer engagement and sales.

Jun 3, 2026 · 3 min read

technical

Microsoft Drops MAI-Thinking-1, Its First Reasoning Model

Microsoft launched MAI-Thinking-1, a 35B-active-parameter reasoning model trained from scratch with no distillation from OpenAI's GPT series. The 1T-parameter MoE hits 97% on AIME 2025 and matches Claude Opus 4.6 on SWE-Bench Pro, signaling Microsoft's bid for model independence.

Jun 3, 2026 · 3 min read

technical

Nvidia's Vera Rubin Enters Full Production at Computex

Nvidia put Vera Rubin into full TSMC 3nm production: 88-core Arm Vera CPU plus next-gen Rubin GPU, with Samsung, SK hynix and Micron all qualified on HBM4. AWS, Google Cloud, Azure and Oracle line up as first cloud customers; broad availability lands in H2 2026.

Jun 3, 2026 · 3 min read

technical

Anthropic Files Confidential IPO With SEC at $47B ARR

Anthropic filed confidentially with the SEC to go public, booking $47B in annualized revenue from Claude subscriptions. The filing follows a $65B Series H that valued the company at $965B — already ahead of OpenAI's $852B private mark.

Jun 3, 2026 · 3 min read

technical

Anthropic Brings Claude Mythos to 150 Critical Infra Orgs

Anthropic expanded Project Glasswing to 150 organizations across 15+ countries, putting Claude Mythos on power, water and healthcare infrastructure. Early runs already surfaced 10,000+ high-severity vulnerabilities, including 400 critical at Cloudflare.

Jun 3, 2026 · 3 min read

technical

Trump's AI Order Sets Up Federal Cybersecurity Clearinghouse

President Trump signed an executive order setting up a federal AI cybersecurity clearinghouse to coordinate vulnerability remediation with industry. The order pushes voluntary partnerships and frames AI as core to US critical-infrastructure defense.

Jun 3, 2026 · 3 min read

Page 1 of 156