The evening briefing.
Today across AI and tech: Nvidia faces Senate scrutiny over China chip sales, Airbnb launches a new AI lab, and a fusion startup raises $465M for a Microsoft power plant.
AI Policy & Geopolitics. Senator Elizabeth Warren has called for Nvidia CEO Jensen Huang to testify before the Senate regarding China AI chip sales, intensifying scrutiny on export controls and data center policy. This move highlights ongoing concerns about the geopolitical implications of advanced AI hardware. Meanwhile, new bipartisan legislation aims to restrict state-level AI regulation in the US, seeking a more unified federal approach.
AI Capabilities & Safety. Anthropic's co-founder warned that AI needs a "brake pedal" to prevent autonomous development, underscoring growing concerns about control and safety. This comes as OpenAI's Codex helped uncover a new HTTP/2 Bomb DoS attack, demonstrating AI's dual-use potential in both discovery and exploitation of vulnerabilities. Separately, the NSA is reportedly using Anthropic's Mythos for cyber attacks, raising questions about state-sponsored AI operations.
AI Product & Infrastructure. Major tech companies are pushing forward with AI integration and infrastructure. Airbnb CEO Brian Chesky plans to launch a new AI lab, indicating a strategic shift towards in-house AI development. Meta is reportedly deploying facial recognition on its smart glasses and is exploring cost-saving measures by building data centers in tents. Google also released Gemma 4 12B, a new model capable of running high-performance AI on standard laptops, expanding on-device AI capabilities.
Emerging Tech & Funding. Quantum computing firm Quantinuum, a Honeywell spin-off, closed flat in its Nasdaq debut after an upsized offering, marking a significant public market entry for the sector. Meanwhile, Helion, a fusion energy startup backed by Sam Altman, secured $465 million to build a power plant for Microsoft by 2028, highlighting substantial investment in advanced energy solutions. These developments showcase continued venture capital interest in high-risk, high-reward technological frontiers.
Senator Warren invites Nvidia CEO Jensen Huang to a Senate hearing on China AI chip sales
Senator Elizabeth Warren is pressing Nvidia CEO Jensen Huang over export controls, China sales, and data-center policy as Congress scrutinizes the AI chip boom. The invitation comes amid growing concerns about the geopolitical implications of advanced AI hardware.
Airbnb’s Brian Chesky plans to launch a new AI lab
Airbnb CEO Brian Chesky announced plans to launch a new AI lab, indicating a strategic shift towards in-house AI development. Chesky previously stated that existing LLM products were not quite ready for partnership.
AI needs a 'brake pedal', warns Anthropic co-founder
Anthropic co-founder Jack Clark warned that AI could reach a point where it develops without human input, emphasizing the need for a 'brake pedal'. This statement highlights growing concerns about AI autonomy and control.
Zcash bug could have let attackers print cryptocurrency out of thin air
A Zcash bug was discovered that could have allowed attackers to create cryptocurrency without authorization. The Zcash Foundation stated there is no evidence of unauthorized value creation.
Filtr is a new privacy tool that blocks ads in almost every iPhone and Mac app
Filtr, a popular ad blocker, now blocks ads within almost all iPhone and Mac apps, including web browsers. This new feature is enabled by the latest Apple software, enhancing user privacy.
Quantinuum closes flat in Nasdaq debut after upsized offering
Honeywell's quantum computing company, Quantinuum, closed flat in its Nasdaq debut following an upsized offering. Honeywell will retain a majority stake and continue as a strategic customer and partner.
The Instagram Plus subscription has officially launched
Instagram Plus, a new optional subscription plan, has launched globally at $3.99/month. Paying users will gain tools for reaching wider or more specific audiences.
NSA using Anthropic's Mythos for cyber attacks
The NSA is reportedly utilizing Anthropic's Mythos model for cyber attacks. This development raises questions about the use of advanced AI in national security operations.
ChatGPT's memory is getting better, especially if you're on the free tier
OpenAI has significantly improved ChatGPT's "dreaming" architecture, enhancing its memory capabilities. These improvements are particularly noticeable for users on the free tier.
OpenAI’s Codex helps discover HTTP/2 Bomb DoS attack that can nuke over 30GB of RAM
OpenAI’s Codex assisted in discovering an HTTP/2 Bomb DoS attack that can consume over 30GB of RAM within seconds, knocking web servers offline. Some major web servers have already released patches for this new attack technique.
New bipartisan legislation takes a big step forward in restricting state regulation of AI
New bipartisan legislation aims to restrict state-level regulation of AI, seeking to establish a more unified federal approach. The draft bill is intended to solicit feedback from stakeholders and experts before its formal introduction.
Meta's ships facial recognition on smart glasses
Meta is reportedly shipping facial recognition capabilities on its smart glasses. This development follows previous reports of Meta exploring facial recognition for its wearable devices.
Google Gemma 4 12B nearly matches 26B benchmarks and runs on your laptop
Google has introduced Gemma 4 12B, a new model designed to bring high-performance, multi-modal intelligence to standard laptops. This model nearly matches the benchmarks of its 26B counterpart, expanding on-device AI capabilities.
Helion, Sam Altman-backed fusion startup, raises $465M to build a power plant for Microsoft
Fusion startup Helion, backed by Sam Altman, has raised $465 million to accelerate its goal of building a power plant for Microsoft by 2028. This fresh infusion of cash will support the company's advanced energy development efforts.