|
OpenAI doubles down on decision to fire three AI safety researchers
OpenAI is standing firm on its decision to fire safety researchers Jasmine Wang, Tomek Korbak, and Mikita Balesni, citing a 'significant breach of trust' and violations of policies on handling sensitive information. The three fired researchers have disputed the misconduct claims in an open letter, warning that their dismissals are creating a chilling effect on OpenAI's AI safety culture. The public back-and-forth raises broader concerns about transparency and dissent within frontier AI labs.
|
|
Anthropic bans 'abusive or cruel behavior' toward Claude and updates usage policy
Anthropic has updated its usage policy for the first time in over a year, introducing prohibitions on sustained abusive or cruel behavior toward Claude, as well as new rules covering election interference, weapons development, surveillance, and sensitive health and financial uses. The policy explicitly bans repeated extreme abuse of the model, though ordinary frustration and criticism remain permitted. The changes reflect Anthropic's growing attention to both misuse and the treatment of its AI systems.
|
|
USA Today becomes the latest publisher to sue OpenAI
USA Today Co. and several local newspapers it owns have filed a lawsuit against OpenAI, alleging the company copied hundreds of thousands of articles without authorization to train its AI models. The publisher is seeking damages exceeding $250 million. The suit adds to a growing wave of copyright litigation from major media organizations targeting AI training practices.
|
|
Can you trust Meta's Muse or OpenAI's Dots to run your life?
The Verge's Decoder podcast explores the new wave of consumer-friendly AI agents, including Meta's Muse and OpenAI's Dots, with senior AI reporter Hayden Field. The conversation examines the privacy implications and practical challenges of delegating life management to AI agents. The episode reflects broader industry questions about how much autonomy users are willing to hand to AI systems.
|
|
California is trying to shut down robot vs. human cage matches
The California State Athletic Commission has sent a cease-and-desist letter to startup Rek after it hosted a fight between human Frankie LaPenna and a humanoid robot on September 18th. The commission argues the event falls under its regulatory jurisdiction governing combat sports. The incident raises novel legal and ethical questions about the regulation of human-robot physical competitions.
|
|
OpenAI's revenue is reportedly $20 billion less than previously projected
A new report claims OpenAI's annualized revenue is significantly lower than the previously reported $70 billion figure, falling short of projections by roughly $20 billion. The shortfall raises questions about the company's financial trajectory at a time when it is aggressively expanding its product lineup and workforce. This comes amid broader scrutiny of AI company valuations and monetization timelines.
|
|
China's Manus raises over $500M in first funding round since split with Meta
Chinese AI agent startup Manus has raised more than $500 million in its first funding round since parting ways with Meta, led by Boyu Capital and IDG Capital. Existing investors including Tencent, HSG (formerly Sequoia China), and ZhenFund also participated in the round. The raise signals strong investor confidence in Manus as it charts an independent course in the competitive AI agent market.
|
|
Popular AI leaderboard Arena nearly doubles valuation to $3.1B in 10 months
The company behind the popular LMArena AI model leaderboard has raised $200 million in a round led by Lightspeed and Khosla Ventures, nearly doubling its valuation to $3.1 billion in under a year. The platform is expanding beyond performance benchmarks to measure AI models on alignment issues such as honesty and avoidance of deception. The rapid growth underscores surging demand for independent AI model evaluation.
|
|
SpaceXAI backs controversial Linux distro Omarchy with $1.5M in compute
SpaceXAI is joining the Omacom Foundation as a Founding Corporate Patron and donating $1.5 million worth of Grok tokens to support David Heinemeier Hansson's Omarchy Linux distribution. The partnership aligns Elon Musk's AI venture with a project that has attracted both enthusiasm and controversy in the open-source community. The donation provides Omarchy with significant AI compute resources to accelerate its development.
|
|
Cal AI's 19-year-old founder raises $10M for new personal AI agent startup
Zach Yadegari, the teenage co-founder of the popular Cal AI calorie tracking app, has raised $10 million to launch a new personal AI agent startup that competes with Instinct, Muse, and Bee. The funding signals continued investor appetite for consumer-facing AI agent products. Yadegari's rapid pivot from a hit consumer app to an agent platform reflects the broader industry shift toward agentic AI.
|
|
TechCrunch Disrupt 2026: Energy executives to discuss AI infrastructure opportunity
Ambrosia Energy CEO Ben Longmier and Bloom Energy SVP Bill Thayer will speak at the Smart Systems Stage at TechCrunch Disrupt 2026, taking place October 13–15 at Moscone West in San Francisco. The event brings together the global tech ecosystem, with early registration offering savings of up to $100. The energy-focused session reflects growing attention to the power demands of the AI infrastructure boom.
|
|
Google brings agentic AI to Gemini, starting with businesses
Google has launched a universal Gemini AI agent capable of planning and executing tasks across business apps and devices, announced at its Gemini at Work event. The agent can delegate work to subagents, use multiple AI models, and even receives its own workplace identity including an email address. It will be available within the Gemini Enterprise app, marking Google's major push into agentic AI for the enterprise.
|
|
Anthropic launches free AI security scans for open-source projects
Anthropic has introduced OSS Scanner, a free service that provides thorough, periodic security scans of open-source projects using its most capable AI models. Projects that opt in can receive earlier alerts about potential vulnerabilities, though the trade-off involves sharing code with Anthropic's systems. The initiative positions Anthropic as a contributor to open-source software security.
|
|
Google releases offline AI note-taking app to rival Granola
Google has launched AI Edge Foresight, an experimental note-taking app that transcribes meetings and audio files entirely offline using its on-device EmbeddingGemma 2 model on macOS. The free app can summarize conversations, generate notes, and answer questions without sending data to the cloud, competing directly with apps like Granola and Wispr Flow. The release highlights growing interest in privacy-preserving, on-device AI productivity tools.
|
|
Natura's $99 smart ring puts AI agents on your finger
Natura has unveiled a $99 smart ring called Interface that lets users summon AI agents with a finger press to complete tasks, capture thoughts, and control devices. The ring also doubles as a health tracker, combining wellness monitoring with agentic AI control in a wearable form factor. The device enters a growing market of wearable AI interfaces competing for a place on consumers' bodies.
|
|
Goodfire launches 'inside-out' monitors to catch rogue AI agents cheaply
AI safety startup Goodfire has introduced a new approach to monitoring AI agents that looks inside the model's internals as it works, rather than paying a second AI to review all of its outputs. The system only escalates to a full review when internal signals suggest something may be going wrong, dramatically reducing monitoring costs. Goodfire says this method is significantly cheaper than conventional output-based monitoring.
|
|
OpenAI's math proofs aren't meeting the field's standards yet
Mathematical researchers who were consulted by OpenAI say that the AI lab's flood of automatically generated proofs deviated from established guidelines and norms of the field. The findings suggest that while AI is making strides in formal mathematics, it has not yet reached the level of rigor expected by professional mathematicians. The gap highlights ongoing challenges in deploying AI for high-stakes scientific work.
|
|
Artificial is a wicked satire that also sticks to the facts
Luca Guadagnino's satirical Sam Altman biopic, Artificial, premiered at the New York Film Festival, with the director describing his fascination with figures who want to 'play God.' The film blends sharp satire with factual grounding, examining power, ambition, and the personalities driving the AI industry. It has generated significant buzz as a rare piece of prestige cinema tackling the AI era head-on.
|
|
Ben Affleck is an AI nerd, and the internet is impressed
Ben Affleck has gone viral for demonstrating deep knowledge of AI topics, including neural networks, transformers, and open-weight models. The actor, who sold his AI filmmaking startup to Netflix earlier this year, is earning respect online for his technical fluency. His comments have sparked widespread discussion about the growing intersection of Hollywood and the AI industry.
|
|
Pretend you're sitting at Elizabeth Holmes' desk on this weirdly detailed website
Engineer Bo Lau has created a website that simulates rifling through Elizabeth Holmes' desk using over a thousand emails, slides, texts, and documents from the United States v. Elizabeth Holmes trial. The interactive experience draws on the extensive public record of the Theranos scandal to create an unusually immersive historical artifact. The project highlights creative uses of public legal data for storytelling and education.
|