DeepSeek’s Sequel Set to Extend China’s Reach in Open-Source A.I.; The New York Times, April 24, 2026

Meaghan Tobin and Cade Metz, The New York Times; DeepSeek’s Sequel Set to Extend China’s Reach in Open-Source A.I.

"DeepSeek released its models as open source, which means others can freely use and modify them. By contrast, OpenAI and Anthropic kept their leading models proprietary. The episode demonstrated that an open-source system could perform almost as well as closed versions. In the months that followed, Chinese firms released dozens of other open-source models. By the end of 2025, these models made up a significant share of global A.I. usage.

On Friday, DeepSeek released a preview of V4, its long-awaited follow-up model, which it intends to open source. The new model excels at writing computer code, an increasingly important skill for leading A.I. systems. It significantly outperformed every other open-source system at generating code, according to tests from Vals AI, a company that tracks the performance of A.I. technologies.

DeepSeek released its new model just days after Moonshot AI, another Chinese start-up, introduced its latest open-source model, Kimi 2.6. While these systems trail the coding capabilities of the leading U.S. models from Anthropic and OpenAI, the gap is narrowing.

The implications are meaningful. Using A.I. to write code is faster and frees up human programmers to focus on bigger issues. It also means people can use DeepSeek’s latest release to power A.I. agents, which are personal digital assistants that can use other software applications on behalf of office workers, including spreadsheets, online calendars and email services."

Agency in the Age of AI; Time, April 14, 2026

 John Palfrey , Time; Agency in the Age of AI

"OpenAI’s recent acquisition of OpenClaw, an open-source, autonomous AI agent designed to run locally on a user’s computer, is a sign that AI agents are quickly being given more responsibilities and more access—from emails to bank accounts, a decision with unintended consequences, including deleted inboxes and Amazon Web Services outages. Peter Steinberger, the founder of OpenClaw, said he wants to “build an agent that even my mum can use.” But there is a difference between using technology to improve efficiency and giving technology agency that humans should hold. 

These developments prompt hard questions, particularly for young people who are seeking agency in their personal and professional lives. Does it make sense to train to be an actuary if AI is supposed to be good at predicting unknown outcomes based on data? Is it worth the cost today to train to be a lawyer or an accountant or pursue higher education at all when all the answers are supposedly at our fingertips? Put another way, what does agency look like in an era dominated by the spread of AI?"

AI agents are scrambling power users' brains; Axios, April 4, 2026

 Megan Morrone, Axios ; AI agents are scrambling power users' brains

"A growing number of software developers say AI coding tools are frying their brains. 

The big picture: The most popular agentic AI systems have triggered something that looks a lot like addiction among some of tech's highest performers."

USPTO announces agentic AI-assisted evaluator for patent eligibility determinations; United States Patent and Trademark Office (USPTO), April 1, 2026

United States Patent and Trademark Office (USPTO) ; USPTO announces agentic AI-assisted evaluator for patent eligibility determinations

"As part of the U.S. Patent and Trademark Office's (USPTO) continued efforts to incorporate artificial intelligence (AI) into agency operations—first with the Artificial Intelligence Search Automated Pilot Program, or “ASAP!,” for patent prior art references followed by the Trademark Classification Agentic Codification Tool, or “Class ACT,” for trademark searching—the USPTO today announced the first-of-its-kind agentic AI tool to assist in patent eligibility determinations under 35 U.S.C. §101. 

America’s Innovation Agency’s new AI system, termed “McConaughey Agentic Tasking Technology Helping Examiner Workload,” or “MATTHEW,” for short, will help examiners tackle the thorniest of eligibility questions as to whether claims presented are an abstract idea or a patent-eligible invention. “MATTHEW will greatly enhance our ability to make the close calls—or any call, really—as I herewith also suspend all applicable precedent, including Desjardins, Alice, and Mayo,” said USPTO Director John A. Squires. “Basically, in terms of eligibility, if MATTHEW says your invention is ‘Alright, Alright, Alright,’ then it’s ‘Alright, Alright, Alright’ with the USPTO.” 

“Initially, we had some concerns that we would be introducing a three-part test in place of the two-part test under Alice and Mayo, but I think we’ll be al…um, okay,” he continued.

“We want to equip our examiners—the best in the world at what they do—with the best tools to assist them,” said Director Squires. “In fact, MATTHEW was selected after careful evaluation of best-in-breed offerings, including the ‘Binary Eligibility Engaged Translation Language Environment Joint User Interface Computational Evaluator,’ or ‘BEETLEJUICE,’” he stated. “But the coders had some issues in testing when they said the name three times. I hope they’ll be al…um, okay,” remarked the Director. 

When asked if the USPTO licensed its tool in light of famed actor McConaughey’s recent Name Image and Likeness (NIL) ‘non-traditional’ registrations, Director Squires retorted, “Well, he’s the one who said, ‘trademark yourself!’—I think the Founders would have wanted this.” When asked if he had heard from Mr. McConaughey’s lawyers, Director Squires produced an unintelligible, guttural chanting sound and began rhythmically beating his chest with his fist.

For more information on this trailblazing AI system, please visit the USPTO website."

Autonomous AI Agents Have an Ethics Problem; Undark, March 5, 2026

 ADAM SCHIAVI, Undark; Autonomous AI Agents Have an Ethics Problem

AI-powered digital assistants can do many complex tasks on their own. But who takes responsibility when they cause harm?

"As a bioethicist and specialist in neurointensive care, I deal directly with human moral agency and the essence of personhood when treating patients. As a researcher, I study the use of synthetic personas animating AI agents and their use as stand-ins of human counterparts. Here is the problem that I see: Granting AI personhood, even in limited capacity, risks formalizing the most dangerous escape hatch of the agentic era — what I will call responsibility laundering. This allows us to say, “It wasn’t me. The agent/bot/system did it.”

Personhood should not be about metaphysics or claims about an inner nature. It is a legal and ethical instrument that allocates rights and accountability. It is a social technology for assigning standing, duties, and limits on what can be done to an entity. If we grant personhood to systems that can act persuasively in public while remaining functionally unaccountable, we create a new class of actors whose harms are everyone’s problem but nobody’s fault.

There is a key concept here that we can use from my field, medicine. In clinical ethics, some decisions are justified yet still leave a “moral residue,” a kind of emotional echo or sense of responsibility that persists after the action because no options fully satisfy competing obligations. This residue accumulates over time, causing a “crescendo effect” that occurs even when conscientious clinicians are doing their best inside imperfect systems. That remainder matters because it reveals something basic about moral life, namely that ethics is not only about choosing; it is about owning what remains afterwards."

Can coding agents relicense open source through a “clean room” implementation of code?; Simon Willison's Weblog, March 5, 2026

 Simon Willison's Weblog ; Can coding agents relicense open source through a “clean room” implementation of code?

"Can a model trained on a codebase produce a morally or legally defensible clean-room implementation?"

What Is Agentic A.I., and Would You Trust It to Book a Flight?; The New York Times, November 25, 2025

 Gabe Castro-Root, The New York Times ; What Is Agentic A.I., and Would You Trust It to Book a Flight?

"A bot may soon be booking your vacation.

Millions of travelers already use artificial intelligence to compare options for flights, hotels, rental cars and more. About 30 percent of U.S. travelers say they’re comfortable using A.I. to plan a trip. But these tools are about to take a big step.

Agentic A.I., a rapidly emerging type of artificial intelligence, will be able to find and pay for reservations with limited human involvement, developers say. Companies like Expedia, Google, Kayak and Priceline are experimenting with or rolling out agentic A.I. tools.

Travelers using agentic A.I. would set parameters like dates and a price range for their travel plans, then hand over their credit card information to the bot, which would monitor prices and book on their behalf...

Think of agentic A.I. as a personal assistant, said Shilpa Ranganathan, the chief product officer at Expedia Group, which is developing both generative and agentic A.I. trip-planning tools.

While the more familiar generative A.I. can summarize information and answer questions, agentic tools can carry out tasks. Travelers benefit by deputizing these tools to perform time-consuming chores like tracking flight prices."

The Pentagon is throwing $200 million at ‘Grok for Government’ and other AI companies; Task & Purpose, July 14, 2025

 MATT WHITE , Task & Purpose; The Pentagon is throwing $200 million at ‘Grok for Government’ and other AI companies

"The Pentagon announced Monday it is going to spend almost $1 billion on “agentic AI workflows” from four “frontier AI” companies, including Elon Musk’s xAI, whose flagship Grok appeared to still be declaring itself “MechaHitler” as late as Monday afternoon.

In a press release, the Defense Department’s Chief Digital and Artificial Intelligence Office — or CDAO — said it will cut checks of up to $200 million each to tech giants Anthropic, Google, OpenAI and Musk’s xAI to work on:

• “critical national security challenges;”
• “joint mission essential tasks in our warfighting domain;”
• “DoD use cases.”

The release did not expand on what any of that means or how AI might help. Task & Purpose reached out to the Pentagon for details on what these AI agents may soon be doing and asked specifically if the contracts would include control of live weapons systems or classified information."