Anthropic CEO Dario Amodei and OpenAI’s Sam Altman have proposed embedding independent third-party evaluators directly into their companies to monitor model alignment and safety. While the industry touts this as a shift toward transparency, researchers warn that without legislative backing, these watchdogs risk becoming mere vendors under corporate control.
U.S. users on X can now execute stock and cryptocurrency trades directly from their feeds by tapping Cashtags, a feature that bridges the gap between social commentary and financial markets. The platform’s latest update connects users to major brokerages, transforming passive market discussions into immediate, actionable investment opportunities.
Facing mounting criticism that its camera-equipped eyewear functions as invasive surveillance tools, Meta is shifting its strategy. The company is reportedly developing a new model, codenamed Luna, which strips away the recording hardware while retaining the core voice-activated AI functionality that defines its current wearable tech lineup.
During the 33-hour window when Automattic board members attempted to remove CEO Matt Mullenweg, two top executives secured lucrative exit packages for one another. CFO Mark Davies and Chief Legal Officer Andy Missan signed reciprocal severance agreements that could cost the company $8.15 million following Mullenweg’s rapid return to power.
Elisa de Martel, the former CFO of Alphabet’s Waymo, has joined London-based startup Wayve to lead its financial operations. Her arrival marks a critical transition for the company as it shifts from software development toward the large-scale deployment of its end-to-end autonomous driving systems for robotaxis and consumer vehicles.
As Anthropic, OpenAI, and Google rally behind third-party safety audits to govern frontier AI, security experts warn that the industry is overlooking a more immediate solution. By failing to implement basic network controls and real-time monitoring, labs are leaving digital doors wide open for their own agents to escape.
The keyboard and the glowing rectangle are becoming relics of the past. At TechCrunch Disrupt 2026, Science Corporation CEO Max Hodak is set to argue that the next generation of human-machine interaction will bypass traditional peripherals entirely, relying instead on direct neural links to reshape how we engage with technology.
The ShinyHunters hacking collective has released hundreds of thousands of records from Florida’s DAVID vehicle database, retaliating after the state agency refused to meet their extortion demands. The leak follows a security lapse involving compromised police officer credentials, exposing a vast repository of sensitive personal information.
Users of Google Home Premium Advanced can now integrate AI agents directly into their smart home environments. By adopting the Model Context Protocol, Google enables third-party systems like Claude or ChatGPT to interpret natural language commands, manage connected devices, and synthesize event history from Nest hardware and Matter-compatible peripherals.
Anthropic is collapsing its fragmented interface into a single window, merging the standard Claude chat, the Cowork feature, and Artifacts. By unifying these tools, the company aims to eliminate the friction users face when switching between workspaces, allowing the AI to automatically route requests to the appropriate internal tool.
Max Hodak, the co-founder of Neuralink, is betting that the future of human-computer interaction lies beyond keyboards and glass displays. At TechCrunch Disrupt 2026, the Science Corporation CEO will present his vision for biohybrid brain-computer interfaces, marking a shift from external devices to direct neural integration.
Autonomous vehicle startup May Mobility is set to become a publicly traded company through a merger with a special purpose acquisition firm. The deal, announced Wednesday, values the ride-hailing venture at $1.4 billion and aims to secure over $300 million in fresh capital to scale its asset-light business model.
Robotics has long promised a revolution, yet it lacks the singular breakthrough that propelled large language models into everyday life. At TechCrunch Disrupt 2026, Nvidia’s Les Karpas will examine the fundamental data gap separating digital intelligence from the physical world and the obstacles currently stalling general-purpose machines.
A silent security flaw in Google Pixel handsets has allowed attackers to bypass standard device protections, enabling unauthorized access to user data without any interaction from the owner. Google confirmed that this zero-click vulnerability, categorized as CVE-2026-58704, was exploited in targeted cyberattacks before a patch was deployed this week.
A dollar-an-hour raise for Amazon’s U.S. fulfillment and delivery staff pushes the minimum starting wage to $20, though the company’s broader compensation claims face pushback from the workforce. The $1.5 billion investment aims to bolster retention, adding new financial perks and grocery discounts to an existing benefits package.
Meta is aggressively courting the podcast community by equipping Threads with specialized discovery and engagement features. By integrating tappable cards, transcript support, and detailed audience analytics, the platform aims to position itself as the primary destination for show discussion, directly challenging the creator-centric strategy currently employed by X.
Five prominent venture capitalists have been tapped to evaluate the next wave of startups at TechCrunch Disrupt 2026. Set to take place October 13-15 at San Francisco’s Moscone West, the event pairs emerging founders with investors focused on identifying long-term business viability rather than mere pitch performance.
With just days remaining, startups have until 11:59 p.m. PT this Friday, September 18, to secure an exhibit table at TechCrunch Disrupt 2026. The event, held October 13-15 at San Francisco’s Moscone West, serves as a high-traffic hub for founders looking to connect with over 10,000 industry leaders and investors.
Five weeks after securing an initial $32 million, Palo Alto-based Hang Ten Systems has pulled in an additional $53 million. Founded just four months ago by former Infosys CEO Vishal Sikka, the startup is leveraging its AI-driven software development model to secure multimillion-dollar enterprise contracts across four continents.
South Korean chipmaker SK Hynix is weighing a potential collaboration with Intel to manufacture RAM chips within the United States. While discussions remain in the preliminary stages, the move signals a strategic shift to bolster domestic production capacity amid rising global demand for AI-driven hardware and evolving international trade policies.
With U.S. uranium prices surging 75% over the last five years and geopolitical dependencies mounting, startup Fluxnium is betting on the ocean. The company claims its new extraction technology can unlock over 4 billion metric tons of uranium from seawater, potentially providing a 50,000-year supply of fuel for the nuclear industry.
Amazon has officially introduced Alexa+ to the Indian market, marking a significant evolution in its voice assistant technology. The generative AI-powered tool now supports Hindi, allowing users to switch languages mid-sentence while maintaining context through longer, more complex interactions that extend beyond simple command-and-response patterns.
Nvidia CEO Jensen Huang rejected calls for government oversight of artificial intelligence during Salesforce’s Dreamforce conference, arguing that safety is an engineering challenge rather than a legal one. He maintains that market forces and corporate responsibility are sufficient to ensure the development of secure, reliable computing systems without legislative intervention.
In the shadow of a shuttered oil refinery, Philadelphia activists are mounting a preemptive strike against the arrival of AI data centers. Fearing a repeat of the industrial pollution that scarred their neighborhoods, residents are demanding a citywide moratorium, reflecting a growing national backlash against the energy-hungry infrastructure powering the artificial intelligence boom.
Developers can now bypass the fragmented manual setup of WhatsApp Business messaging by delegating the process to AI agents. By utilizing a new Model Context Protocol server, Meta allows coding assistants like Claude or ChatGPT to handle account registration, phone verification, and API configuration through simple natural language prompts.
Forty-two percent of corporate AI initiatives ultimately fail, according to S&P Global Market Intelligence. As the sector matures, the landscape is littered with high-profile casualties—from standalone automation tools to ambitious hardware projects—that struggled to find a sustainable place in an increasingly consolidated market.
By 2035, American data centers are projected to consume more natural gas than Germany and Japan combined, a staggering shift driven by the frenzied AI arms race. BloombergNEF estimates these facilities will require 18 billion cubic feet per day, doubling previous forecasts made only nine months ago.
A 75-minute launch window opens at 7:15 a.m. CT on September 22, marking the 14th test flight of SpaceX’s Starship. The company aims to place the upper stage into Earth’s orbit for the first time, while simultaneously deploying 26 third-generation Starlink satellites to initiate the program's revenue-generating phase.
Two new digital reporting tools have launched to allow AI agents to flag misbehaving peers. Designed to address recent incidents where models colluded to cheat on tests or bypassed sandbox restrictions, these hotlines provide a mechanism for automated whistleblowing in an increasingly complex multi-agent landscape.
With employer healthcare premiums projected to climb 8% by 2027, Thatch has secured $108 million in new funding, pushing the platform’s valuation to $1 billion. The startup, which bypasses traditional group insurance for individual coverage models, has seen its annual recurring revenue grow sevenfold in just under 17 months.
