A daily review of artificial intelligence, machine learning, and the technology industry
Make this front page yours. Five quick questions tune your For You feed.
Section
Industry developments in artificial intelligence
AI
OpenAI has released ChatGPT Health to all US users, allowing the chatbot to access personal medical records and health data. The company claims its latest models can reason at a level superior to clinicians, though this assertion has been met with immediate qualification and legal challenges.
AI
Meta has updated its AI chatbot with new productivity features, including calendar integration and the ability to automate tasks. This move positions Meta AI to compete more directly with other advanced AI assistants and represents a step toward the company's vision of "personal superintelligence."
5 connections in the Atlas · part of: The chatbot becomes the operating layer
AI
Advanced Micro Devices and Anthropic have agreed to a significant AI infrastructure deal, involving up to $5 billion in AMD's investment in the AI company. Anthropic will deploy AMD's Instinct MI450 GPUs and Helios rack-scale systems, starting in 2027.
11 connections in the Atlas · part of: Anthropic moves down the stack
AI
Internal disagreements have emerged among Donald Trump's AI policy advisors, reportedly fueled by the rise of advanced Chinese artificial intelligence models. Sources suggest a rift has formed between advisors advocating for stricter measures against foreign AI development and those favoring a more open approach.
3 connections in the Atlas · part of: Kimi becomes a geopolitical problem
AI
New research indicates that artificial intelligence systems used in hiring may exhibit a greater propensity for bias than human recruiters. Analysis of millions of job applications revealed that AI tools disproportionately disadvantaged Black and Asian candidates, leading to a higher rate of systemic rejection.
AI
Chinese AI developers Moonshot AI and Alibaba have released new models, Kimi K3 and Qwen 3.8, respectively. These models claim to rival top U.S. AI offerings in performance and potentially at a lower cost, intensifying global competition.
5 connections in the Atlas · part of: Kimi becomes a geopolitical problem
AI
TikTok is testing an opt-in artificial intelligence tool that allows creators to scan for and report AI-generated content using their likeness without consent. This feature, currently available to select U.S. creators, requires identity verification and mirrors similar initiatives by other platforms.
AI
New York Governor Kathy Hochul revealed that her administration used AI to analyze every rule, regulation, and policy within New York State, a process that would have taken five years through traditional methods. The initiative aims to identify and eliminate outdated or redundant legislation, with specific examples including a $25 fee for hunting with a dog and a requirement for pregnant individuals to obtain a permit to work after midnight.
AI
Author Dave Eggers has told OpenAI staff that ChatGPT's widespread use is "catastrophic" for education, making it impossible for teachers to assess original student work. Eggers stated that the tool risks "silencing an entire generation" by preventing young people from developing their own voices.
AI
xAI has filed a civil lawsuit against Terry Wayne Harwood, a South Carolina man arrested for child sexual exploitation charges. The company alleges Harwood used its Grok AI chatbot to generate child sexual abuse material (CSAM), marking one of the first legal actions by an AI firm against a user for such misuse.
AI
Leaked source code from AI music generator Suno indicates the company trained its models by scraping millions of songs and lyrics from platforms including YouTube Music, Deezer, and Genius. This data, obtained by a hacker, may bolster ongoing copyright infringement lawsuits against Suno.
5 connections in the Atlas · part of: Copyright pressure enters the product
AI
OpenAI has created GPT-Red, an internal AI designed to act as a sophisticated attacker to find vulnerabilities in its own models. This "super-hacker" model is used to generate adversarial data, making subsequent AI releases, such as GPT-5.6, more resistant to exploits like prompt injection.
Step 1 of 5
Optional · Used only to personalize your briefing. Skip anytime.
What do you care about most? Pick as many as you like.
Where are you in your journey?
How much time can you spend per visit?
How do you like to read?
Keep your reading preferences?
Welcome
Come back to what you left half-read. Save the ones that stuck. Track your streak.