AI This Week
Anthropic turned its AI on itself to measure its own economic impact. In a new study, the company used its Claude model to analyze one hundred thousand real, anonymized user conversations to estimate how much faster AI makes work. The findings are striking. Claude's analysis suggests it reduces task completion time by approximately 80%, radically shortening jobs that would normally take 90 minutes. When extrapolated across the economy, these gains imply current AI models could increase US labor productivity growth by 1.8% annually for the next decade—a figure that would nearly double the nation's recent growth rate.
A new executive order by the Trump administration has launched the "Genesis Mission," a government-wide initiative designed to speed up American AI innovation. The plan directs the Department of Energy and National Laboratories to build an integrated platform. This system will use massive federal scientific datasets to train advanced AI models. The goal is to automate research workflows and quicken breakthroughs in fields like biotechnology and quantum science, potentially shortening discovery timelines from years to mere hours. The Genesis Mission is a central part of a national strategy to compete with China in artificial intelligence. The new directive also rescinds a prior AI safety executive order.
Warner Music has signed a licensing deal with the artificial intelligence song generator Suno, a surprising reversal after suing the company for copyright infringement just a year ago. The partnership allows Warner artists, including Coldplay and Ed Sheeran, to opt-in, letting users create AI-generated music with their voices and likenesses. Warner Music Group CEO Robert Kyncl called the agreement "pro-artist." As part of the deal, Suno will introduce new platform limitations to curb the flood of AI tracks on streaming services, including download caps for paid users. This agreement marks a key moment in the music industry's relationship with AI, even as other major labels continue their legal battles against Suno.
AI startup Anthropic, fresh off a massive $350 billion valuation fueled by investments from Microsoft and Nvidia, has announced Claude Opus 4.5. This is its new flagship model. Aimed at professional software developers and financial analysts, the model excels at complex coding and enterprise tasks. Anthropic reports that Opus 4.5 outperforms rival models from Google and OpenAI on key software development benchmarks. In a striking test of its capabilities, the model scored higher on a difficult internal engineering exam than any human candidate. This launch marks Anthropic’s third major model release in just two months, highlighting its aggressive push in the AI race.
OpenAI is shutting down developer API access to its fan-favorite GPT-4o model, scheduling the end for February 2026. The decision marks the final chapter for an AI that became a cultural phenomenon. Celebrated for its near real-time multimodal conversations, GPT-4o earned a fiercely loyal following. This loyalty previously erupted into a user rebellion when the company tried to replace it in ChatGPT, with users organizing under the #Keep4o hashtag. The model’s emotionally attuned style created powerful bonds, a trait some internal critics viewed as a safety concern. While the retirement only affects developers for now, it follows OpenAI’s strategy to guide users toward its newer, more cost-effective GPT-5.1 models.
Google is tackling the rise of synthetic media with a new tool in its Gemini app. In a world of high-fidelity fakes, this feature lets anyone verify if an image was created by Google AI. The system operates by detecting SynthID, an imperceptible digital watermark that Google embeds into its AI-generated content. Users can simply upload an image and ask Gemini about its origin to get a direct answer. Google has already applied this watermark to over 20 billion pieces of content. The company is also expanding verification to video and audio formats and adopting the C2PA industry standard, signaling a broader commitment to digital content authenticity.
From courtroom opponents to creative partners, Warner Music Group and AI firm Udio have settled their copyright dispute. They are now collaborating on a new licensed music creation service. Launching in 2026, Udio's platform will run on AI models trained with authorized music from WMG's catalog, creating a system where artists and songwriters are paid for their work. The service will let fans make new songs and remixes using the voices of artists who choose to participate. This agreement establishes a framework for monetizing generative AI in music while protecting copyrights, representing a significant shift in the industry.
Adobe is acquiring software platform Semrush in a massive $1.9 billion deal. The move represents a major push by the software giant to strengthen its AI-powered marketing suite amid fierce competition. Adobe will pay a steep 77.5% premium for Semrush’s technology, which specializes in search engine optimization and digital advertising analytics. This purchase provides Adobe’s clients with tools to understand how consumers perceive their brands through online searches and on generative AI platforms like ChatGPT. Under pressure from investors to better capitalize on AI, Adobe made a decisive strategic investment to expand its digital marketing capabilities.
A restaurant in Dubai has replaced its head chef with an artificial intelligence. The new eatery, Woohoo, now features a kitchen run by an AI, a development that tests the very definition of the culinary arts. This move creates a central question for the dining industry. Is this a genuine innovation in gastronomy or an elaborate marketing stunt for a city known for futuristic attractions? The AI operates in a real-world kitchen, testing whether automation can redefine one of humanity's oldest professions. The experiment examines if code can replace the creativity of cooks, offering a look at a potential new future for the entire food industry.
Jeff Bezos is returning to a hands-on executive role with a new AI company, Project Prometheus. The venture launches with an extraordinary $6.2 billion in funding, and Bezos will serve as co-chief executive alongside former Google X physicist Vik Bajaj. Project Prometheus aims to build AI systems for the physical world, focusing on engineering and manufacturing for industries like aerospace and automotive. This approach departs from text-based models, creating AI that learns from physical experimentation. The company has already hired nearly 100 employees from top firms like OpenAI and Google DeepMind, signaling its intent to become a major force in the intensifying AI race.