As AI systems become more capable and autonomous in high-stakes settings like health, science, education, and coding, they will need to remain helpful, honest, transparent, and safe in situations they have not seen before. This requires generalizing to new contexts, new pressures, longer and more complex interactions, and across domains that differ from those seen during training.

We find that reinforcement learning on realistic scenarios targeting beneficial traits can produce broad improvements across dozens of benchmarks measuring aligned and beneficial behavior. These alignment gains generalize beyond the domains used for training and persist under adversarial pressure. — Read More

What do the Uber budget blowout, a 24x token multiplier, and context teach us about building a real business case for AI Agents in production?

Let’s start in April 2026, when Uber’s CTO Praveen Neppalli Naga said something every engineering and finance leader building or buying agentic AI should sit with: “I’m back to the drawing board, because the budget I thought I would need is blown away already.”

Claude Code adoption jumped from 32% to 84% of Uber’s 5,000-engineer org between December 2025 and March 2026. By April, the entire annual AI budget was gone. … According to Gartner’s March 2026 analysis, agentic models require between 5 and 30 times more tokens per task than a standard chatbot. — Read More

More than three and a half years after ChatGPT’s initial release, AI assistants are now used by millions of people worldwide, and the competitive landscape is changing fast. While OpenAI’s chatbot is still the most popular assistant globally, its market share has dipped below 50% for the first time as users are migrating between different assistants like Google’s Gemini, Anthropic’s Claude, and xAI’s Grok, according to analytics firm Sensor Tower’s State of AI Report for 2026.

ChatGPT’s growth has been impressive. It became the fastest app ever to reach 1 billion monthly users, as Sensor Tower reported this month. Notably, OpenAI counts weekly active users, and it last reported 900 million of them in February. The chatbot still remains the most popular AI assistant worldwide with over 1.1 billion monthly users, followed by Gemini with 662 million and Claude with 245 million. — Read More

We are excited to introduce Sakana Marlin, our first commercial product—an autonomous research assistant for business, built on our long-horizon reasoning technology. Give it a research topic, and Marlin works autonomously for up to roughly eight hours, crafting a detailed strategy report up to a hundred pages long, along with executive summary slides.

Sakana Marlin is designed to take on the kind of substantial strategy research that a Chief Strategy Officer (CSO) and a small team might otherwise spend weeks on. — Read More

The accelerating build-out of edge AI is starting to redefine how people interact with AI, shifting the focus from massive global data mining and analysis in huge AI data centers to faster results, greater efficiency, and much more targeted workloads at the edge.

In both cases, the emphasis is still on processing and moving data at blazingly fast speeds. But at the edge, there is less data to process, and the distances that data has to travel are shorter. Hyperscalers emphasize contextual search, massive simulations, and training of large language models. At the edge, goal may be as limited as feeding commands to a robot about how much pressure is needed to pick up an object, or telling a car to jam on the brakes because a pedestrian just darted across the street. Small language models that are domain- and workload-specific replace more generalized capabilities in LLMs.

There is demand for both, but as the edge takes shape, it is beginning to look very different from what OpenAI or Anthropic does. — Read More

Prometheus, the physical AI startup co-founded by Jeff Bezos and Vik Bajaj, the former co-founder of Verily, Google’s life sciences unit, announced it raised $12 billion at a $41 billion valuation.

… Prometheus is building what it calls an “artificial general engineer” — software capable of automating the design and manufacturing of complex physical systems, from jet engines to drug compounds. — Read More

AI has made execution extremely cheap. What’s expensive is committing to the wrong thing. I’ve been thinking of this as Selection Debt: the cost of moving quickly on a poor foundational premise.

… As Derek Sivers puts it, you have to say no to many things to leave space for the one or two HELL YESes. — Read More

Anthropic recently published a policy paper about AI, jobs, and what governments should do if AI causes major labor-market disruption. The original paper is here: Anthropic’s economic policy proposal.

The paper is about the question: if AI creates huge wealth while also replacing a lot of human labor, who gets the money, who pays the costs, and who gets blamed?

Anthropic is warning that AI could seriously disrupt jobs. It says governments should prepare now with better unemployment systems, wage support, retraining, public benefits, and possibly new taxes or wealth-sharing mechanisms later. That sounds responsible. It also protects Anthropic’s business interests. — Read More

We’re more productive than ever. AI allows us to generate code at supersonic speeds, unfold entire modules in seconds, and ship thousands of lines of code. It’s easier to pick up tasks and generate value, even in unfamiliar codebases. But there’s a dark side. AI-assisted code generation isn’t free; there’s a hidden cost that we as an industry are only beginning to realize: AI burnout. Are we dangerously ignorant to this problem? And how can we cope with it?

Across the industry, developer voices are rising, flagging the troubling state they find themselves in: increasing fatigue, a constant race to keep up with the ever-rising pace of work, mixed feelings about AI-assisted coding, and a persistent energy drain. Vibe-coding turns into doom-coding. — Read More

From a pure AI perspective, nothing Apple showcased during their WWDC keynote yesterday was particularly groundbreaking. In fact, much of it featured capabilities long since available in other AI tools and services – in some cases, years ago. And guess what? That doesn’t matter. Based on what we saw yesterday, Apple is set to win in AI. At least from a consumer perspective.

I know how crazy this sounds. It’s not just that Apple has been viewed as behind in AI for the past few years, it’s that they’ve been more or less a laughingstock given how they tried to roll out ‘Apple Intelligence’ two years ago and failed to the point of settling lawsuits around false advertising. But if Apple is actually able to roll out what they showcased yesterday – I’ll get to the caveats below – and there’s reason to believe they can this time, they’re about to infuriate many people and companies across a wide swath of industries. That’s because Apple seems on the verge of doing what they always do: watching new products and services come about and then jumping in later with a better user experience to win the day. — Read More