AI Pulse: Daily Digest — April 29, 2026
Summaries are AI-generated. Click through to read the original reporting.
Elon Musk began his testimony in the lawsuit he filed against OpenAI CEO Sam Altman and co-founder Greg Brockman, alleging the company abandoned its nonprofit mission in favor of profit. Observers noted Musk appeared flat and underprepared on the stand, leaning heavily on a personal narrative about wanting to save humanity rather than presenting sharp legal arguments. The trial's outcome could determine whether OpenAI is permitted to operate as a for-profit company ahead of its anticipated IPO.
Read more →OpenAI has renegotiated its agreement with Microsoft, ending the exclusivity clause that had previously locked its models to Azure — in exchange, Microsoft receives a larger revenue-sharing arrangement. The amended deal clears the path for OpenAI to distribute its models through competing cloud platforms, most immediately Amazon Web Services. Amazon wasted no time, announcing a slate of OpenAI model offerings on AWS Bedrock, including a new agent service, just one day after the deal was confirmed.
Read more →Google has entered a classified agreement with the U.S. Department of Defense allowing the military to use its AI models for any lawful government purpose, according to The Information. The deal came just days after Anthropic refused a similar DoD request, citing concerns about domestic mass surveillance and autonomous weapons applications. The agreement sparked internal protests at Google, with employees demanding CEO Sundar Pichai block the Pentagon from accessing its AI systems.
Read more →Chinese regulators have ordered Meta to unwind its multibillion-dollar acquisition of Manus, the AI agent startup, following a months-long government probe. The decision is a significant blow to Mark Zuckerberg's push into agentic AI and illustrates how deepening U.S.-China tech rivalry is complicating cross-border deals involving AI companies with Chinese roots. The case highlights the growing difficulty for AI founders in severing ties with China when seeking Western investment or acquisition.
Read more →Ineffable Intelligence, a British AI lab founded just months ago by former DeepMind researcher David Silver — the architect behind AlphaGo — has secured $1.1 billion in funding at a $5.1 billion valuation. The company's core ambition is to build AI systems capable of learning entirely from self-generated experience, bypassing the need for human-labeled training data. The massive early-stage raise signals intense investor appetite for research bets on more autonomous, self-improving AI.
Read more →GitHub has announced it will move Copilot to a consumption-based billing model, citing unsustainable "escalating inference costs" driven by its heaviest users. Under the new structure, developers will pay based on how much AI they actually use rather than a flat subscription fee. The shift reflects a broader industry reckoning with the economics of AI tooling, as providers struggle to balance accessibility with the real cost of running large models at scale.
Read more →Anthropic has launched a set of "creative connectors" that allow Claude to integrate directly with popular professional tools including Adobe Creative Cloud, Blender, Ableton, Affinity, and Autodesk. The move is part of Anthropic's push into the creative industry, following the recent launch of Claude Design, and positions Claude as an AI collaborator embedded within existing creative workflows rather than a standalone chat interface. The connectors allow Claude to read context from and take actions within these applications in real time.
Read more →NVIDIA has unveiled Nemotron 3 Nano Omni, an open multimodal model that unifies vision, audio, and language processing into a single system — eliminating the latency and context loss that occurs when AI agents pass data between separate specialized models. NVIDIA claims the architecture enables up to 9x more efficient AI agents, with faster and more coherent responses across complex, multi-modal tasks. The model is designed for deployment in agentic systems that need to process documents, audio, and video simultaneously.
Read more →European regulators have told Google it must open Android to competing AI assistants, arguing that Gemini's preferential default placement constitutes an unfair competitive advantage. Google pushed back, calling the intervention "unwarranted" and warning it could harm the user experience on its platform. The ruling, if enforced, could set a significant precedent for how AI assistants are distributed on mobile operating systems across the EU.
Read more →Industry analyst Ming-Chi Kuo reports that OpenAI is developing a smartphone in collaboration with MediaTek, Qualcomm, and Luxshare, designed around AI agents that would replace traditional apps as the primary interface. The device would represent a fundamental rethinking of the smartphone paradigm, with agents handling tasks that users currently navigate through discrete applications. The rumored phone would complement OpenAI's previously reported plans for AI-powered earbuds.
Read more →Taylor Swift has filed a series of trademark applications in an escalating effort to protect her likeness, voice, and brand from AI-generated imitations. Swift has been one of the most high-profile targets of AI deepfakes and voice cloning, and her legal strategy reflects a growing trend of celebrities turning to intellectual property law to combat synthetic replicas. Legal experts caution, however, that trademark law is an imperfect tool for this problem and the effort may face significant hurdles in court.
Read more →Japan Airlines has launched a pilot program deploying humanoid robots to handle luggage sorting at Haneda Airport, with an eye toward eventually using them for cargo loading and aircraft cabin cleaning. The trial is driven by Japan's acute labor shortage, which has made automation an increasingly urgent priority across service industries. If successful, the program could accelerate the adoption of humanoid robots in airport and logistics operations globally.
Read more →As tech companies race to build AI infrastructure across the United States, many rural communities are mounting organized resistance to data center construction in their backyards. Concerns center on water consumption, strain on local power grids, noise pollution, and the limited number of permanent jobs these facilities actually create for local residents. The backlash is creating a growing geographic and political divide over where — and whether — the physical backbone of the AI economy should be built.
Read more →A new wave of low-skill hackers is leveraging AI tools to execute sophisticated cyberattacks that previously required deep technical expertise, effectively democratizing offensive capabilities in dangerous ways. The trend was on display at DARPA's AI Cyber Challenge last August, where AI systems scanned 54 million lines of code for vulnerabilities — tools that, in the wrong hands, could be weaponized at scale. Cybersecurity researchers warn that the same AI advances driving productivity gains are simultaneously lowering the barrier to entry for malicious actors.
Read more →Chinese AI firm DeepSeek has released a preview of V4, its next flagship model, which can process significantly longer prompts than its predecessors and demonstrates notable improvements in reasoning and instruction-following. The release reinforces DeepSeek's position as a serious global competitor in frontier AI development, continuing to challenge the assumption that cutting-edge models require Western compute and capital. MIT Technology Review highlights the release as one of three reasons the model matters for the broader trajectory of AI development.
Read more →Despite enormous investment and widespread adoption of AI tools, many enterprises are still struggling to convert AI deployments into measurable financial returns — a gap MIT Technology Review describes as the missing step between hype and profit. The analysis points to fragmented data infrastructure, unclear ownership of AI initiatives, and the difficulty of scaling pilots into production as the primary culprits. The piece argues that without solving these foundational problems, the current wave of AI investment risks repeating the pattern of previous technology cycles that promised more than they delivered.
Read more →Summaries are AI-generated. Click through to read the original reporting.