Plus: OpenAI has unveiled its long-awaited super app. The AI firm Anthropic has got the clearest glimpse yet at what's really going on inside large language models as they answer questions or carry out tasks. What they found ranges from the mundane to the unnerving. Researchers at the company built a tool called the Jacobian lens (or J-lens) and used it to uncover a hidden area, which they named the J-space, inside its flagship LLM, Claude. The J-space contains words related to the response a model is working on but may not ultimately produce. If Claude were a person (which it is not), you might say these hidden words reveal what's on its mind before it actually speaks.

PCWorld explores strategic AI delegation, emphasizing three key questions to determine which tasks are appropriate for AI assistance versus human decision-making. The guide matters for productivity-focused users seeking to leverage AI chatbots like ChatGPT, Claude, and Gemini for tedious data compilation and repetitive tasks while avoiding over-reliance. Key recommendations include delegating boring, repeatable tasks like email management and voice memo transcription to AI, but retaining all final decision-making authority for yourself. You know what's harder than making decisions?

The AI firm Anthropic has developed a technique that has given it the clearest glimpse yet at what's really going on inside large language models as they answer questions or carry out tasks. What they found ranges from the mundane to the unnerving. Researchers at the company built a tool called the Jacobian lens (or J-lens) and used it to uncover a hidden area, which they named the J-space, inside Claude Opus 4.6, a version of Anthropic's flagship LLM released in February.

Experiments in using AI to build AI show that the future doesn't just belong to the frontier labs. These days, the frontier AI labs are all racing to build self-improving models . Some believe it's the surest route to superintelligence --as AI improves itself in a mind-melting loop, the thinking goes, it will eventually surpass human comprehension (and perhaps even control). That's all well and good, but I have a newsletter to produce. I wondered if recursive self-improvement might also be useful for me.

Now you can direct Anthropic's Claude Cowork AI from your phone Now you can direct Anthropic's Claude Cowork AI from your phone Max subscribers get first access. At the start of the year, Anthropic debuted Claude Cowork, an offshoot of its coding agent that gives people a way to ask Claude for help with tasks on their computer. After expanding the capabilities of the software in March, Anthropic is giving users a way to manage Cowork from their phones. Once you update to the latest version of the Claude app on Android and iOS, look for the new Cowork tab in the sidebar. To be clear, today's update doesn't mean you can use Anthropic's agent to automate tasks on your phone.

PCWorld explores the risks and benefits of using AI agents like Claude for email management, following Notion Mail's recent shutdown that left users dependent on AI sorting. AI email automation offers appealing benefits including reduced inbox clutter and better organization, but poses operational risks like misfiling or accidental deletion. Privacy concerns remain significant as AI agents access sensitive personal and financial data, requiring users to carefully weigh convenience against potential security risks. When I learned that Notion, the popular online workspace service, was shutting down its Notion Mail product, it wasn't the shutdown itself that got my attention. No, it was this: because so many Notion users had handed over their email sorting duties to AI agents, they'd stopped bothering to open their inboxes . Letting AI agents sort through all your email has long been considered a killer app for AI, although the convenience doesn't come without some serious risks.

Let's start with a game. Open up your chatbot of choice--Claude, ChatGPT, Gemini--and type "Give me a random number between 1 and 10." You're going to get 7. Almost always. Now type "Another" and you'll get 3 or 4. Type "Another" again and you'll get 8 or 9. That won't work every time--but if it did for you, you may wonder if I have superpowers.

A researcher found that using Anthropic's Claude Opus 4.7, he could break into the website of Front Gate--used by every festival from Lollapalooza to Bonnaroo--and freely issue any ticket he chose. Fears about AI tools capable of autonomous hacking usually involve nightmare scenarios like the theft of nuclear launch codes or zeroed-out bank reserves. Far more plausible, it turns out, is asking AI to gain super-administrator access on a ticketing website and then issuing yourself and all of your friends free VIP backstage passes to Bonnaroo. That was the discovery of security researcher Ian Carroll, who used the AI tool Claude Opus 4.7 in April to discover a technique that allowed him full access to the systems of Front Gate Tickets, which handles ticketing for practically every major US music festival, from Lollapalooza and South by Southwest to Austin City Limits. Carroll found that Front Gate, which like Ticketmaster is a subsidiary of the event company Live Nation Entertainment, had a bug in its website that he--with Claude's help--could exploit to gain access to millions of customer or staff records and freely issue tickets for any event, of any value, to himself or whoever he chose.

PCWorld explores how human writing can exhibit AI-like characteristics, using Claude Sonnet 4.6 to analyze writing patterns for potential AI tells. The analysis identified parenthetical asides as the top AI-like trait, followed by em dash overuse and long sentences exceeding 35 words. Despite detecting these patterns, Claude rated the human writing as mostly authentic with only a 3/10 AI similarity score, suggesting stylistic quirks shouldn't be mistaken for artificial intelligence. What could be better than an em dash? They pepper my writing, adding dramatic pauses and emphasizing my biggest points.

PCWorld explores how human writing can exhibit AI-like characteristics, with an author using Claude Sonnet 4.6 to analyze their own content for artificial intelligence tells. Parenthetical asides emerged as the top AI indicator with 67 instances across 11,700 words, followed by em dash overuse and filler words like "actually." Despite these patterns, Claude rated the writing as mostly human, scoring only 3 out of 10 on AI detection due to strong personal voice and varied sentence structures. What could be better than an em dash? They pepper my writing, adding dramatic pauses and emphasizing my biggest points.