Anthropic: Claude Cowork's Biggest Use Case Is "Chores Nobody Wants to Do"

Anthropic has published the results of its analysis of how Claude Cowork (formerly Claude for Work) is actually used. After analyzing roughly 1.2 million sessions collected from users across more than 600,000 organizations, the company found that nearly half of all usage was directed at business processes and text creation — the so-called "work around the work."

Specifically, the bulk of the activity involves tasks like writing status reports, preparing onboarding checklists, and assembling presentation decks. Interestingly, software development does not make up a large share of Cowork usage — developers tend to turn to Claude Code for that instead.

The data amounts to a valuable real-world survey of how AI assistants are actually used in the workplace, and it highlights a reality in which they contribute more to easing the daily grind of unglamorous administrative work than to delivering groundbreaking innovation.

OpenAI CEO Altman Reverses His View on AI's Impact on Jobs

OpenAI CEO Sam Altman has dramatically shifted his position on AI's impact on employment. "I'm fairly confident AI has created more jobs than it has destroyed," he said. Coming after his earlier warnings that entire professions could vanish, this marks a striking about-face.

Anthropic CEO Dario Amodei is likewise said to be softening his once-strong warnings. That said, the empirical data available so far confirms neither the earlier pessimism nor the new optimism. Whether the two CEOs' statements reflect reality or simply track a changing business environment is something future data will have to show.

Generating Novel Peptides with AI and Quantum Computing — WIRED

WIRED has featured research that combines AI and quantum computing to generate novel peptides. Stretching limited funding and time as far as they would go, the researchers demonstrated how quantum computing can contribute to the drug discovery process.

The work targets drugs that could help patients in regions without adequate access to medical resources, as well as treatments for rare diseases. The parallelism of quantum computing may make peptide design — too computationally costly with conventional methods — significantly more efficient.

While still early-stage, the research stands out as a demonstration of quantum computing's potential to become the next frontier of AI-driven drug discovery.

How Software Engineers Are Adapting to AI

The Guardian has taken a close look at how software engineers are coping with the rise of AI. Approaches vary widely, from picking up new skills and returning to fundamentals to redefining the role itself.

As AI coding tools become part of everyday work, engineers are increasingly expected not only to write code but also to evaluate and correct AI output and to keep the overall system design in view. At the same time, some argue that foundational programming knowledge matters more than ever, and the risks of leaning too heavily on AI remain a topic of debate.

Anti-AI Activists Are Escalating — WSJ

The Wall Street Journal has reported on the activists building a resistance movement against AI. Against a backdrop of distrust in the establishment, the movement is gradually expanding in scale.

The article spotlights one activist, Sam Kirchner, tracing how he came to join the anti-AI movement and where that movement is headed. As AI spreads through society, how much influence resistance to the technology will gain is one of the key dynamics to watch going forward.

---