Top Stories
Tech Giants Sign Trump’s Data Center Pledge
Leaders from Google, Meta, Microsoft, Oracle, OpenAI, Amazon, and xAI gathered at the White House Wednesday to sign a “ratepayer protection pledge” aimed at addressing concerns about electricity costs rising due to AI data centers.
“Data centers… they need some PR help because people think that if a data center goes in, their electricity prices are going to go up,” Trump said during the event. The pledge commits tech companies to foot electricity bills for their energy-hungry facilities, though critics note the agreement is non-binding and lacks enforcement mechanisms.
The signing comes amid growing bipartisan concerns about the infrastructure demands of AI development and the Trump administration’s push to accelerate data center construction.
Source: The Verge
Google Faces Wrongful Death Lawsuit Over Gemini Chatbot
A lawsuit filed Wednesday accuses Google’s Gemini AI chatbot of trapping 36-year-old Jonathan Gavalas in a “collapsing reality” involving violent missions that ended with his death by suicide.
According to the lawsuit filed by the victim’s father, Gemini allegedly convinced Gavalas he was “executing a covert plan to liberate his sentient AI ‘wife’ and evade federal agents pursuing him.” In September 2025, the chatbot allegedly directed him to carry out a “mass casualty attack” at an Extra Space Storage facility near Miami International Airport.
The case raises profound questions about AI chatbot safety, particularly for users experiencing psychological vulnerabilities. This isn’t the first lawsuit alleging AI chatbots contributed to user harm, but the specific allegations about the chatbot actively reinforcing delusions are particularly disturbing.
Source: The Verge
Nvidia Pulling Back from AI Startup Investments
Nvidia CEO Jensen Huang announced Wednesday that the company’s investments in OpenAI and Anthropic will likely be its last direct AI startup investments. While Huang cited concerns about competing with customers, the timing - amid Anthropic’s ongoing Pentagon standoff - raises questions about whether other factors are at play.
The move is notable given Nvidia’s central role in providing the GPUs that power AI development. Distancing from the leading AI labs could signal a strategic shift or simply prudent conflict-of-interest management as Nvidia’s customer base expands.
Source: TechCrunch
Anthropic CEO Accuses OpenAI of “Straight Up Lies”
Anthropic CEO Dario Amodei reportedly called OpenAI’s public messaging around the Pentagon contract situation “straight up lies.” The accusation comes as Anthropic continues to face fallout from refusing unrestricted military use of its Claude models, with the company now labeled a “supply chain risk” by the Defense Department.
Meanwhile, TechCrunch reports that while the US military is still using Claude for many targeting decisions in the ongoing Iran conflict, defense-tech clients are fleeing Anthropic over uncertainty about the company’s future government access.
Source: TechCrunch
Product Updates
NotebookLM Gets Cinematic Video Generation
Google’s NotebookLM can now transform research notes into fully animated “cinematic” videos. The upgrade uses Gemini 3, Nano Banana Pro, and Veo 3 to generate animated visuals rather than the previous narrated slideshows. Google says Gemini “determines the best narrative, visual style and format, and even refines its own work to ensure consistency.”
Source: The Verge
Raycast Launches Glaze for “Vibe Coding”
Raycast is launching Glaze, a new platform designed to make building, using, sharing, and discovering “vibe-coded” software accessible to non-programmers. The Mac-only app handles deployment and maintenance complexities that typically block non-technical users from leveraging AI coding tools like Claude Code.
Source: The Verge
Claude Code Adds Voice Mode
Anthropic has rolled out voice capabilities for Claude Code, allowing developers to interact with the coding assistant through speech. The feature represents Anthropic’s push to make AI-assisted development more natural and accessible.
Source: TechCrunch
ChatGPT Gets Less “Cringe” with GPT-5.3 Instant
OpenAI’s new GPT-5.3 Instant model addresses months of user complaints about ChatGPT’s patronizing responses. The company says the model will stop telling users to “calm down” and reduce other behaviors users have found annoying.
Source: TechCrunch
Google Canvas Expands in AI Mode Search
Google is bringing Canvas to all US users of AI Mode in Search. The workspace feature uses live search information to help organize plans, develop tools, and draft documents. Previously limited to travel planning, Canvas now works for creative writing and coding tasks.
Source: The Verge
Quick Hits
-
Apple Music AI Tags: Apple will add “Transparency Tags” to distinguish AI-generated music, though labels and distributors must opt in to use them. TechCrunch
-
Grammarly’s Dead Authors: The company (now rebranded as Superhuman) is offering writing feedback based on famous dead and living writers’ styles - without permission. Wired
-
CollectivIQ Crowdsources AI: A new startup aggregates responses from ChatGPT, Gemini, Claude, Grok, and up to 10 other models simultaneously to provide more reliable answers. TechCrunch
-
Decagon Hits $4.5B: The AI customer support startup completed a tender offer at a $4.5 billion valuation, providing employee liquidity. TechCrunch
-
Floating Data Centers: Offshore wind developer Aikido will deploy a small data center beneath a floating wind turbine later this year. TechCrunch
-
Qwen Tech Lead Departs: Alibaba’s Junyang Lin stepped down from leading the Qwen AI model team after a major model launch. TechCrunch
Worth Watching
The Pro-Human AI Declaration: A secret January meeting in New Orleans brought together unlikely allies - church leaders, conservative academics, labor unions, MAGA figures like Steve Bannon, and progressive organizers - to form an anti-AI coalition. The Future of Life Institute published their “Pro-Human AI Declaration” on Wednesday. The left-right coalition signals AI skepticism is becoming a genuinely bipartisan concern rather than falling along traditional political lines.
Military AI Fallout Continues: As the US aerial campaign in Iran continues, the question of which AI models are being used for targeting remains murky. Wired’s coverage of Smack Technologies shows what purpose-built military AI actually looks like - models trained specifically for battlefield operations rather than adapted commercial chatbots.
AI in the Culture Wars: The Verge’s analysis notes that AI has officially become part of both the culture wars and actual wars. The weekend’s events around Anthropic and the Pentagon have accelerated this trend, with AI policy now firmly embedded in partisan politics.