An eighth-grader just put three major AIs through an… · M&A Beginners 🎓
| View this email in your browser |
![]() Models & Agents for BeginnersAI explained simply — for beginners and teens.
|
🎧 Today's episode Episode 206 · An eighth-grader just put three major AIs through an empathy test and found they often missed the mark. 2026-10-24 ▶ Listen now |
The Big StoryAn eighth-grader named Ellianna Gu from Salt Lake City decided to test how well the biggest AI chatbots understand teenage emotions. She created 84 different emotional scenarios that real teens might face and asked ChatGPT, Claude, and Gemini to respond to each one. The results showed that the AIs often gave replies that felt flat or missed the emotional point entirely. Think of it like texting a friend who always answers with the safest, most generic reply instead of really hearing how you feel. In one case the AI might suggest a quick fix when the person just needed someone to listen. In another the response might acknowledge the feeling in one sentence but then immediately pivot to advice instead of staying with the emotion. This matters because lots of students already turn to these tools for homework help, advice, or just to talk through a rough day. When the responses lack warmth, it can leave people feeling more alone instead of supported. The test covered everyday situations like dealing with friendship drama, pressure from school, family arguments, and moments of self-doubt that many eighth-graders recognize. Ellianna found that even when the AIs got the facts right, the tone often felt off or overly formal. For you, it means the AI you use for school projects or late-night questions might not always pick up on the tone you’re looking for. The good news is that noticing these gaps helps everyone push for better tools. You can read the full story of her experiment on the Times of India site linked below and think about what kinds of emotional questions you would test next. Her project shows that even a middle-school student can run a meaningful check on how these systems handle real human feelings. That kind of curiosity is exactly what pushes the technology forward. Source: timesofindia.indiatimes.com Explain Like I'm 14You know how when you’re texting a friend, your phone guesses the next word based on what you’ve already typed and what people usually say in that kind of message? Now picture the same idea but stretched across millions of conversations at once. The AI looks at the words you just sent, checks patterns it has seen in billions of other chats, and picks the reply that feels most likely to fit. It doesn’t have feelings of its own, so it’s really just choosing the safest-sounding next chunk of text. When the situation involves strong emotions, that “safest” choice can land flat because the model is trying to avoid saying anything risky. The result is an answer that sounds polite but doesn’t quite match the mood you were hoping for. Ellianna’s test found so many replies that felt off even when the facts were correct because the system is trained to stay neutral and helpful rather than emotionally present. It has no memory of what it feels like to be disappointed or excited, so it can only copy the shape of caring responses it has seen before. That copying works well for simple questions but breaks down when the conversation needs real timing and tone. The model is essentially playing a very advanced game of “what would most people say next” without ever feeling the moment itself. That is why the replies in her 84 scenarios often came across as helpful in theory but distant in practice. Cool Stuff & Try ThisFree AI chat apps you can open right now Six different chatbot apps are compared in one handy guide, showing which ones let you start chatting for free and which paid plans start at about eight dollars a month. The list includes well-known names plus a couple of newer options that focus on speed or privacy. You can see exactly how each app handles the same prompt and decide which style feels most natural to you. If you’re curious what each one feels like, you can open the article and click straight through to the free versions. Try asking the same question in two different apps and notice how the tone changes. One might give short, direct answers while another adds more detail or asks follow-up questions. No coding or sign-up is required for the basic free tiers on most of them. You can run the test on your phone during a study break and see which one matches the way you like to talk. The comparison also notes how quickly each app responds and whether it keeps the conversation flowing without repeating itself. That kind of side-by-side check helps you pick the right tool for different tasks like brainstorming ideas or getting quick explanations. Source: tech-insider.org ChatGPT’s voice mode is acting a little robotic lately Some users have noticed that the live voice chat starts sounding normal for the first few words, then speeds up and starts reading like an instruction manual. The change makes the voice feel less like a friend and more like a machine reciting lines. If you already use voice mode for quick questions or brainstorming, you can test it yourself by opening the ChatGPT app on your phone and switching to voice. Try describing a fun weekend plan out loud and see whether the replies stay natural or start racing ahead. It’s a small reminder that even voice AI is still learning how to sound human. The effect seems to happen more often during longer exchanges, which is why some people switch back to typing when they want a slower, more thoughtful conversation. You can compare the two modes in the same app and decide which one feels better for different moments. The observation comes from everyday users who were simply trying to have casual chats and noticed the shift in delivery. Source: reddit.com Quick BitsMemory limit messages suddenly disappeared For about a week ChatGPT kept telling people to start a new chat after roughly ten messages because of a memory limit. The warning vanished just as quickly as it appeared, leaving some users frustrated and others simply relieved. One person said they had already started moving their longer conversations to another chatbot because the constant reminders made chatting feel broken. Then the messages stopped without any announcement, so the same chats that felt impossible a few days earlier now continue without interruption. The sudden change shows how these systems can shift from one update to the next, sometimes fixing a problem before most people even understand what caused it. Source: reddit.com |
💬 Reply to this email — Patrick reads every one. Share: X · LinkedIn · WhatsApp Forwarded this email? Subscribe here — it's free. |
📺 Watch on YouTube · 📝 Read the blog · 🖼 Free image gallery (CC BY-SA) · 📊 Data Hub & Story Trackers · 🧭 Start Here Nerra Network · AI-narrated voice (Grok TTS) · Editorial by Patrick You're receiving this because you subscribed to Models & Agents for Beginners on nerranetwork.com. |
| Issue #206 · Models & Agents for Beginners · Oct 24, 2026 |
