Claude is exceptionally good at inferring test scenarios, but in terms of implementation IMHO it writes tests that are far too much low level and focused on implementation details.

In my existing projects I almost always have my own test patterns and utilities that effectively provides a mini-framework at the abstraction level I consider appropriate. I can tell Claude to use these as basis for new tests and it helps immensely in getting the kind of test implementations I want.

Note though that the reason most of my projects got these mini-test-frameworks was originally to avoid other human developers implementing tests using the same anti-patterns, so it's not a new problem nor an AI coding agent specific one IMHO.

A modern coding agent can have huge context windows. I use Claude Code with Opus and 1M tokens window at work and that's plenty for "normal" coding activities. The real limiting factor are the API token costs.

As of carefully hand-holding the AI, it's often not necessary anymore since a modern coding agent can infer if there is ambiguity and proactively ask the user for further information and make use of structured development methodologies, e.g. TDD with small incremental changes.

Last week I didn't prompt anything except "implement issue xxx" and Claude Code connected to our JIRA, read the issue in question, asked clarifying question, created a plan and submitted it to me for review. I iterated on the plan a couple times as there were parts I didn't like, then let Claude proceed with the implementation. The result was correct except for a minor GUI issue and a performance optimization.

Instant SEPA support for retail payments is still fragmentary and relies on private financial institutions.

The idea with the Digital Euro is that it would effectively have "legal tender" status. The EU would definitely push for widespread adoption.

Completely cashless is likely not going to happen any time soon, but more and more people pay cashless most of the time. I don't even remember the last time I paid something with cash myself.

The idea is not to replace or compete with SEPA: it's to try to replicate some nice aspects of cash in a more and more cash-less society especially for day-to-day transactions.

A cash-less payment currently needs to involve a private financial institution somewhere in the transaction. This is because the central bank only issues cash. This "digital euro" supposedly will allow for cash-less transaction only involving government institutions instead of private ones.

Of course the question is whether you trust more a government institution than a private one and whether this will open the door to an eventual phasing-out of cash down the road...

The summary also mentions him being able to talk with his daughter. Furthermore, being productive is a very important aspect for many people and while it might not be their whole purpose in life, it's still typically very important for their self-realization.

Also a professional setting is usually focused at delivering results and that the technology is successful in that setting is indicative of how advanced it has become.

The SVP is the most "right-wing" party among the major parties in Switzerland. Although firmly "right-wing" I don't think it should be considered "far-right" though.

Note that Switzerland has a number of "major" parties that cover the political spectrum. Even as the largest, the SVP has "only" 30% of the seats. The second largest party has 20% and it's the left-wing SDP. Neither the left parties nor the right parties have absolute majority. The system functions by finding consensus.

IMHO the problem is that many professional are sleeping on AI. They don't take the time to try it and become proficient with it. Part of it is likely due to inertia, part of it due to prejudice or bad experience with earlier iterations.

You can and Claude is pretty good at auto-configuring itself. You still need to tell him what you need though, which means having familiarity over what MPC servers, skills and plugins can do.

I used Gemini. My prompt was: "create a bullet list in html syntax of skills that might be useful or required to use an AI in a professional setting". This is the result:

• Prompt Engineering: Crafting precise, context-rich, and structured prompts to elicit high-quality, accurate responses from generative AI models.
• Critical Thinking & Fact-Checking: Verifying AI-generated outputs for biases, logical fallacies, or "hallucinations" before implementing them in business decisions.
• Data Literacy: Understanding how data is collected, structured, and utilized by AI systems, as well as recognizing data privacy and security boundaries.
• AI Tool Familiarity: Navigating and mastering domain-specific AI platforms (e.g., large language models, AI image generators, automated data analysis tools, and coding assistants).
• Ethical Awareness & Compliance: Understanding the ethical implications of AI use, including copyright laws, intellectual property rights, and company compliance policies.
• Workflow Automation: Integrating AI tools into existing workflows using APIs, low-code/no-code platforms, or automation software (like Zapier or Make) to increase efficiency.
• Iterative Problem Solving: Approaching AI tasks with an experimental mindset—refining, testing, and tweaking inputs continuously to achieve the desired outcome.

As a software engineer working daily with coding agents and AI workflows, I can confirm the list is pretty accurate although of course not exhaustive. IMHO the most important skill is being aware of what an AI can accomplish, which nowadays is a lot.

I'm sure they will, to conduct some undoubtedly ethical reverse engineering.

Because their situation might be different. I do personal stuff on my work devices from time to time but my employer explicitly allows minor personal use and in my jurisdiction they are actually prohibited by law from performing behavioural monitoring unless they have a very specific reason.

It doesn't matter who initiates the download. By your rationale putting compromised packages in package repositories would never be a crime since "the user initiates the transmission", which is obviously not how the courts see it.

The author decided to shut up and consult a lawyer for a reason.

The author made the files available on distribution channels. GitHub Releases very likely qualify as distribution channel. Maven Central 100% qualifies as distribution channel. That is most likely more than enough to be considered "knowingly causes the transmission" as interpreted by the courts.

Artifacts containing the code in questions were made available. That is very likely to qualify as "knowingly causes the transmission" even only as GitHub Release and it 100% qualifies as deployment on Maven Central.