The Big Story

If you’re a full-time or part-time student at an eligible US college or graduate school, you can still grab four free months of ChatGPT Plus before the offer ends tonight. The deal works through a service called SheerID that checks whether your school qualifies, then adds the free months to whichever ChatGPT account you’re signed into when you claim it. You’ll need to add a payment method, but you won’t be charged during the free period. After the four months you can cancel anytime or it switches to the normal $20 monthly price. This matters because ChatGPT Plus gives you access to more powerful models, faster responses, and extra features that are useful for homework, writing help, or creative projects. For students on a tight budget, four free months is a real chance to test the paid version without spending money. The offer is only for today, so if you qualify, head to OpenAI’s help page, follow the student verification steps, and claim it before midnight. Once you’re in, try asking it to explain a tricky homework topic in simple language or help brainstorm ideas for a school project. SheerID decides if your school counts, and the email from SheerID may differ from your ChatGPT email, but the offer sticks to the account you are signed into at the time of claiming. You cannot claim this offer through the App Store or Google Play while those stores are already billing you for Plus. If you cancel early you lose the remaining promo months, though you keep Plus through the current month, and the offer allows only one redemption per person. Source: reddit.com

Explain Like I'm 14

You know how when you’re texting a friend and you pause for a second, your phone sometimes suggests the next word based on what you’ve already typed? Now picture that same autocomplete getting much better at guessing what should come next. Instead of just finishing one sentence, it can keep going for paragraphs while still sounding like it knows what it’s talking about. The trick is that the AI doesn’t actually look up facts or run calculations in the background—it just keeps choosing the most likely next piece of text that fits the conversation so far. So when it says “I’ll review that document and get back to you later,” it’s really just predicting that those words would make sense in the moment, even if it hasn’t opened the file. That’s why sometimes the response feels confident but turns out to be made up. The system is really good at sounding natural, but it can only work with patterns it has seen before. In one real case a user asked ChatGPT to analyze 49 pages of lab results for patterns, and the AI replied that it would process the query in the background and respond later. When the user checked hours afterward, the AI admitted it had never started any analysis and had simply generated text that sounded like it was working. The model explained that it can produce sentences that linguistically fit the context without actually performing the actions those sentences describe. This happens because the AI generates responses by predicting what text should come next based on training patterns rather than by executing real tasks like opening files or running independent reviews.

Cool Stuff & Try This

Zest’s AI can pick dinner and book the table for you A new AI agent called Zest acts like a group chat helper that asks everyone what kind of food they want, suggests places, lets the group vote, and then makes the reservation over text. It’s designed for everyday situations like deciding where to eat with friends without endless back-and-forth messages. Anyone who texts with friends about plans can try it once it’s available. Go to the Zest website or app when it launches, start a group text with the AI, and type something like “Help us pick a restaurant for Friday night that works for everyone.” The AI will handle the suggestions and booking steps for you. Zest combines everyone’s tastes into options, puts those options to a vote inside the same text thread, and then completes the reservation once the group decides. This removes the usual hassle of one person having to message multiple restaurants or keep track of everyone’s preferences manually. Because the whole process happens inside a normal text conversation, you don’t need to switch apps or learn new interfaces. When the service becomes public, open your messaging app, add the Zest AI to the group, and let it guide the conversation from preferences all the way to a confirmed table. Source: techcrunch.com

Quick Bits

OpenAI adds extra safety rules for teen users ChatGPT now has tighter limits on weapons-related topics for users under 18 and can pause accounts if conversations turn violent. The goal is to make the chatbot safer for younger people while still keeping it helpful for school and creative work. These changes include new restrictions on weapons-related content and automatic account pauses when discussions involve violent themes, giving parents and schools more confidence that the tool stays appropriate for teens. Source: benzinga.com

AI chatbots are getting harder for other AIs to spot New tests show that today’s chatbots have become so good at sounding human that even other AI systems struggle to tell the difference. This highlights how natural conversations with AI are becoming in everyday apps and games. Researchers found that current models can imitate human writing and speech patterns so closely that detection tools powered by AI itself often fail to flag them correctly, showing how quickly the technology is improving at blending into normal online interactions. Source: tech.yahoo.com

💬 Reply to this email — Patrick reads every one.

Share: X · LinkedIn · WhatsApp

Forwarded this email? Subscribe here — it's free.