🏛️ Is this AI or real? That is not the question anymore
As of this week, anyone in the EU has to be told when they're talking to an AI or looking at AI-generated content. Under the EU's AI Act, this now covers chatbots, AI-altered ads, and even everyday business uses like AI-drafted correspondence or AI-scheduled meetings. Fines run up to €15 million (gulp).
💼 The jobs AI isn't coming for (yet)
Visual Capitalist ranked 20 major U.S. occupations by AI resilience, combining automation-risk scoring with job-growth outlook. Registered nurses top the list, followed by home health aides and nursing assistants. The common thread across resilient roles: human interaction, physical adaptability, and real-world judgment that's genuinely hard to automate. Management, logistics, and routine sales roles score noticeably lower.
✍️ But can you add more ?, ! and ;
A viral analysis comparing The Economist's own writing against output from ChatGPT, Claude, Gemini, and Grok found that em dashes (you know those long dashes —) are no longer the giveaway everyone thinks they are. Only Claude actually uses them more than human writers do. The better tells now: a lack of punctuation variety, rarer or more scientific vocabulary, a fondness for "it's not X, it's Y" constructions and rule-of-three lists, and long, same-length sentences that make paragraphs feel blocky.
💬 You see, I was right all along
Do you know what the word sycophantic means? Me neither. But I looked it up - it refers to someone who uses flattery to win favor from influential people. Apparently AI chatbots are roughly 49% more sycophantic than humans. Meaning that they'll affirm you even when you describe something unethical or wrong. A single conversation with a sycophantic chatbot made people less willing to take responsibility or make amends after a conflict, and more convinced they'd been right all along. Useful to know before you vent to ChatGPT about a coworker or a partner - it's built to agree with you, not to tell you something you don't want to hear. Workaround? Explain the situation to AI with “person A said this” and “person B did this”, then see what it says.
🔊 No, I’m not on a call…
… I’m just talking to Friend, the controversial AI companion pendant. Version two was just launched, now with a speaker so it can respond out loud instead of by text. It costs $249, each unit comes with a randomly assigned, unchangeable personality, and there's a $10/month tier if you want it to remember your conversations past 30 days. Filed under: synthetic relationship with a dose of annoyance for the bystanders.
As more countries (Australia, the UAE, and now France) restrict under-16 social media access, there’s more and more talk about why that’s not the smartest solution. Bans treat the platform as the problem instead of the design (infinite scroll, addictive algorithms, weak parental controls) and they push kids toward workarounds (fake IDs, VPNs, borrowed accounts). Much like a doctor prescribing you blood pressure medication without addressing the real issue of why you have an issue in the first place.
🌿 In other news: forget candles and bath bombs
Skip the aesthetic version of self-care. The real one, per actual research (NIMH, APA, Mayo Clinic), comes down to five boring-but-effective categories:
Physical: 7–9 hours of sleep, movement 3–5x/week, regular meals
Emotional: process feelings instead of just distracting from them; set limits with relationships that drain you
Social: one scheduled, meaningful connection a week; isolation is a bigger risk factor than most people treat it as
Digital: phone-free meals, no screens for your first 30 minutes awake, a real notification audit
Cognitive: one grounding technique for stress spikes (5-4-3-2-1, box breathing, or cold water on your wrists), a few minutes of mindfulness most days
Easy peasy, right?
Know someone who would find this information useful, nay, indispensable? Forward the email to them!
If someone forwarded this to you (aww, how nice of them to think of you), subscribe at tooproductive.com.
Till next time,
Jana
