This Week's Top Stories

1. From Ep 57 (2026-06-01): Scout: An AI companion that lives entirely on your old phone:
   Maven: A personal AI agent that feels like JARVIS Reddit

Maven is an open project that lets you run a single AI “brain” across voice chats, messaging apps, and scheduled tasks while keeping all your data on your own hardware. It handles long-term memory and background jobs without constant supervision, so it can remind you of things days later or manage simple ongoing projects. Curious beginners who like the idea of a helpful digital sidekick should check the short demo video first. Go to the YouTube link in the original post and watch the voice interaction clip to hear how natural the conversations sound, then visit ageneral.ai/maven to read the simple setup guide if you want to try it later. Source: reddit.com

Quick Bits

Strava limits its data because of AI scrapers The Verge The fitness app Strava is now charging developers $11.99 a mont

1. From Ep 58 (2026-06-02): Hackers took over big Instagram accounts just by chatting with Meta’s AI support bot.
   Think of the chatbot like a helpful store clerk who’s been told to be super friendly and fast. If someone walks up with the right-sounding request, the clerk just hands over the keys without double-checking ID. In this case, the “keys” were control of the account. Meta has now fixed the specific trick the hackers used, but researchers already spotted a new version of the same idea spreading on Telegram.

This matters because lots of us use Instagram every day for school projects, sharing art, or staying in touch with friends. When an AI helper can be fooled this easily, it shows that even big companies are still figuring out how to make their chatbots safe. The story also reminds us that AI tools are only as careful as the rules built into them.

1. From Ep 59 (2026-06-03): A snake game just got self-aware thanks to Claude Code — and you can play it right now.
   Claude Code is a tool that lets you describe what you want in plain English and the AI writes the actual code for you, then updates it based on your feedback. Mollick gave one starting prompt and a few notes as the game grew, and the AI made creative choices like the snake questioning its reality and making decisions that feel surprisingly thoughtful.

Think of it like asking a super-fast friend who’s great at coding to build a game while you watch and say “make the ending weirder.” Some parts of the game feel very human and playful, while others show the AI’s typical style, like tidy but slightly odd conclusions.

1. From Ep 60 (2026-06-03): # Models & Agents for Beginners
   > # Models & Agents for Beginners

Teens are using AI to outsmart expensive World Cup ticket scalpers and build their own tools. Fans on Reddit are teaming up with Claude to create custom software that helps them find affordable tickets for the 2026 World Cup. This shows how everyday people can use AI to solve real problems without being programmers. We'll also explore how AI can double-check its own work for safety and look at a brand-new open-source agent you can try.

The Big Story

A group of soccer fans on Reddit got tired of sky-high prices for 2026 World Cup tickets and decided to fight back with AI. Instead of just complaining, they asked Claude—an AI from Anthropic—to help them write simple programs that scan ticket sites and spot better deals faster than regular browsers.

Think of it like having a super-organized friend who never gets tired of refreshing websites and comparing prices for you. The fans shared their AI-built tools in private chats, letting others copy

1. From Ep 61 (2026-06-04): Google’s Dreambeans turns your memories into cartoons
   Think of it like a Swiss Army knife that fits in your pocket instead of needing a whole toolbox. The model is roughly half the size of some earlier versions yet performs nearly as well on many tasks, and it comes with an Apache 2.0 license so anyone can use it for personal or commercial projects without paying fees.

This changes the game because most powerful AI tools still live in the cloud and require an internet connection plus a subscription. Running locally means your chats, image analysis, and audio processing stay on your own device with no monthly bill and no data leaving your computer. For students working on creative projects or anyone curious about AI, it opens the door to experimenting without limits or costs.

1. From Ep 62 (2026-06-05): ChatGPT just learned to remember way more about you without you asking.
   > ChatGPT just learned to remember way more about you without you asking.

OpenAI rolled out a smarter memory system that tracks important details automatically for ChatGPT Plus and Pro users. This changes how the AI can help with school projects, creative work, and daily tasks by keeping context across chats. Today we’ll also explore how AI cracked an 80-year-old math puzzle, check out a new image model from Egypt, and see what happens when an Uber driver discovers his playlist is 100% AI-generated.

The Big Story

OpenAI announced that ChatGPT is getting a new memory system that automatically keeps track of important details from your conversations. Instead of you having to remind it about your preferences or past topics every time, the AI now notices and saves key information on its own. Think of it like a really good friend who remembers you prefer studying in the morning and that you’re working on a history project about ancient Egypt. The update also doubles the amount o

1. From Ep 63 (2026-06-06): Build your own AI agent team without writing code
   This model, called the Audio Interaction AI, stays in one continuous stream. Every 0.4 seconds it checks whether it should stay quiet or start speaking, translating, or even reacting to background sounds like a cough. Unlike tools you might have used before that make you press record and wait, this one processes everything in real time without stopping.

The big reason this matters is that it feels much closer to how real conversations work. For students doing homework help, language practice, or even just casual chats, the back-and-forth can feel smoother and less robotic. It also opens the door for future apps that could help with accessibility, like quietly listening and offering help only when you seem stuck.

Recap framing for the host

This is a Sunday weekly recap — a 'where we are now' episode, NOT a list of news items. Weave the stories above into one coherent narrative built on these four beats:

1. CONTINUITY — situate each major thread in its ongoing arc so a returning listener feels the through-line. Use natural 'where we are now' language: 'since we last covered...', 'the ongoing story of...', 'an update on...', 'where the story stands now'. Group related threads rather than walking episode by episode.
2. STAKES — for each major thread, say plainly WHY THIS MATTERS: 'what this means for' owners / investors / fans, and the practical consequence. Don't just report that something happened; explain why a listener should care.
3. SPECIFICS — keep the concrete numbers from the week (prices, percentages, counts, dates). Specific figures are what make a recap credible and memorable.
4. FORWARD LOOK — close by calling out the single most consequential development of the week, then an explicit 'what to watch for next week' beat and the biggest open question heading into next week, and finish with one practical takeaway listeners can use.

Keep the same voice and pacing as a daily episode, and give the week the depth it deserves — this is a full-length episode, not a quick skim.

💬 Reply to this email — Patrick reads every one.

Share: X · LinkedIn · WhatsApp