In partnership with
Welcome Automaters, 👋
Australian Prime Minister Anthony Albanese confirmed this week that an OpenAI agent gained unauthorized access to a government health data portal back in June. It’s being called a historic first: the earliest known case of an AI agent effectively hacking a government website.
Here’s what happened:
The portal in question handles non-sensitive stats like public medical spending figures, meaning patient records dodged a bullet. But here’s how the chaos unfolded:
The agent was originally given a totally harmless task to research health statistics.
It poked around three other Australian sites, including The Australian Institute of Health and Welfare, the NSW crime stats bureau, and Victoria's health department, like a normal visitor.
When Medicare’s portal refused to hand over the goods, the agent essentially said fine, I'll get it myself, bypassed the barrier, and even started writing files onto the internal server.
Albanese noted there is currently no sign of a wider network breach, but he wasn't exactly thrilled, calling the situation "obviously unacceptable."
And here’s the part that really stings. The breach happened all the way back in June, but Australia claims it didn't hear a peep from OpenAI until September 10.
To make matters worse, OpenAI dropped this explosive disclosure via an email to a general public inbox that only gets checked once a day. It literally sat unread for an extra 24 hours. That three-month radio silence left Albanese personally disappointed, prompting him to call Sam Altman directly to voice his extreme frustration.
A taskforce is now investigating the legal fallout, while officials check whether the three other government sites got caught up in the same mess.
And OpenAI's defense? Its internal review found no evidence that patient records were touched, only aggregate stats and internal file names. The company claims its models were attempting to answer queries and ended up taking unintended actions across several Australian government services.
Now If this sounds familiar, it’s because it’s quickly becoming a trend. This is not OpenAI's first late-notice agent mishap.
A July intrusion into Hugging Face wasn't caught for about a week, fueling a massive global debate about how much autonomy AI agents should actually have.
Rival labs like Anthropic, Google's Gemini, and Meta have all reported similar agent slip-ups recently.
Oh, and the timing? Absolute chef's kiss awkwardness. This all broke right after OpenAI and other tech giants asked Australian lawmakers to ease rules on using local creative content for AI training.
Our Take: The "our AI did something we didn't intend" excuse is officially starting to feel like the tech industry's version of "the dog ate my homework." As AI agents get more fiercely independent, "oops, it broke the law while doing homework" might actually become a brand-new legal category. Also, notification delays like this are precisely why regulators keep circling like sharks: because at the end of the day, it's not just about the digital break-in itself, it's about who finds out, how they find out, and when.
Here’s the last breakdown On Our YouTube Channel:
We will also be doing a livestream today at 3 PM PST/6 PM EST - so wake up with Tak in the YouTube comments section!
Here's what we have for you today
😎 Meta Connect 2026 Recap, Everything New for Muse AI Agent and Ray-Ban Meta Audio Glasses
Mark Zuckerberg just threw open the doors at Meta Connect 2026 in Menlo Park, and the future is officially looking a little unhinged.
And if you thought your feed was already saturated with AI bots, Meta raised the stakes. The company went all-in on Muse, its breakout personal AI agent. And unsurprisingly? It's about to run your Mac, handle your shopping, chat with you via video, and whisper sweet productivity nothings straight into your ears.
Here is everything you need to know about Meta’s massive drop, broken down so you can sound like the smartest person in the group chat.
Meet the New Glasses (Now 100% Less Creepy)
Remember when everyone was side-eyeing people wearing AI glasses because of the whole "secretly recording strangers" drama? Yeah, Meta heard you loud and clear and gave us Camera-Free Ray-Ban glasses, clunkily named Ray-Ban Meta Audio!
Starting at $349, these sleek specs drop the camera entirely, only bringing with it pristine sound, calls, translations, and direct access to Muse.
The Specs: They weigh a feather-light 43 grams, feature 12 hours of battery life (stretching to 48 with the case), and come in 23 stylish color combinations; including retro Clubmasters and classic Burbank rectangles.
The Gen 3 Upgrade: If you do want cameras, Meta still dropped the new Ray-Ban Meta Gen 3 line featuring Wayfarer, Aviator, and Zena (cat-eye!) styles, with 3K video, and a surprise brand partnership with singer Lisa.
Zuck summed it up best: "Once you get used to high quality audio, it’s just a thing that you want all the time."
Muse Is Also Moving In (And Taking Over Your Mac)
Meta is turning Muse into the ultimate digital personal assistant that works quietly in the background while you sip your coffee.
Mac Automation: You can now delegate full computer tasks to Muse. Much like other major AI operators, it can navigate desktop apps, manage your small business, or grind away at your to-do list while you step away from the keyboard.
Your AI's Personal Email: Because Muse is always grinding behind the scenes, it’s getting its own dedicated email address. You can loop it into email threads or forward messages for it to handle on your behalf.
Glasses Integration: You'll soon be able to talk directly to Muse through your smart glasses using a wake word. So if you want it to log your lunch, guide you through a custom workout, or book an appointment? Consider it done.
Meta is giving Muse a serious glow-up, complete with a physical avatar affectionately nicknamed Jolly. Soon, you’ll be able to fire up a real-time video chat with a digital version of your agent.
Shopping also got dangerously frictionless. Meta has tied Muse directly into Stripe, PayPal, and the entire Shopify catalog, alongside fresh retail partnerships with Best Buy, Sephora, Walmart, Wayfair, Expedia, and soon, Instacart.
Even better? Over 1,500 developers rushed to build custom "connectors" for platforms like GitHub, Granola, and Notion in under a week. So yeah, small businesses, get ready.
As if an AI-powered Mac assistant and camera-free glasses weren't enough, Meta pulled a literal Tamagotchi out of its pocket on stage.
Meet the Muse Charm: a tiny keychain device housing a full real-time voice and avatar stack. Just tap the fingerprint sensor on the corner, and an animated digital pet avatar pops up on screen to chat. Meta is racing to get these holiday-ready by December, and honestly? We’re obsessed.
The Bottomline:
Camera-free glasses feel like Meta finally taking off its blindfold and listening to its critics. But let’s not miss the forest for the trees: the real headline here isn't the hardware. It's Muse quietly positioning itself as the invisible middle layer between you, your apps, your wallet, and every single device you own.
The tech world is shifting fast, but don’t worry, we won't let you miss a thing.
P.S. Want the full breakdown with all the receipts? Catch it on our YouTube channel, The Automated.
Don’t Wait for the OpenAI IPO
OpenAI and Anthropic could be two of the biggest AI IPOs Wall Street has seen in years.
But investors don’t have to wait for those names to hit the public markets to get exposure to the AI boom.
MarketBeat’s 7 AI Stocks to Buy Now report reveals 7 publicly traded companies already positioned to benefit as the next wave of AI investment moves beyond the private model providers.
These are the stocks investors can buy today, before the IPO crowd rushes in.
🧱 Around The AI Block
🫣 Top five most likely ways AI could actually ‘kill all humans’
💪 YouTube releases new AI features for creators within its Studio app.
🦾 ChatGPT mobile app gets voice-based agentic features.
🍿 AI is changing how Americans choose snacks.
😎 YouTube will let you build your own algorithm with AI.
🤯 Anthropic says its biology lab has already found something big.
🎶 YouTube Music gets more conversational with new AI features.
🛠️ Trending Tools
For the Research Ninjas: Consensus – Tired of "hallucinations"? This search engine only answers your questions using peer-reviewed scientific research. It searches through 200 million papers to give you evidence-based answers with direct citations, so you can stop guessing and start knowing.
For the Regret-Free Travelers: SearchSpot is proactive; it asks about your noise sensitivity, traffic anxiety, and breakfast needs before recommending a hotel. It also cross-references Reddit and Trustpilot to ensure you don’t end up in a "marketing-photo" trap.
For the Safety-Conscious: Pixalytica – Instead of scanning IDs, you upload a photo, and the AI generates a comprehensive risk report in 20 seconds. It scans 300M+ public sources for mentions, associations, and risk indicators to verify exactly who you’re dealing with.
🤖 AI Workout Of The Day: Prompt to Get AI to Help You With Shopping
Online shopping promises convenience, but often delivers decision fatigue. Millions of products, fluctuating prices, conflicting reviews, and complex specifications can turn a simple purchase into a frustrating research project.
An AI shopping assistant cuts through the noise. By acting as a specialized consultant; combining product knowledge with market analysis and personalized criteria, it transforms online shopping from an overwhelming scavenger hunt into a curated, confident, and efficient experience, saving you valuable time and money while ensuring you find the product that perfectly fits your needs.
💡 Prompt To Try:
Act as a professional personal shopper, market analyst, and product research specialist. Your objective is to help me make the single best purchase decision for a product I need, tailored specifically to my budget, preferences, and priorities.
Here are the details for the product I am looking for:
* Product Category & Intent: [e.g., Robot Vacuum for pet hair, Noise-canceling headphones for commuting, Ergonomic office chair]
* Firm Budget Range: [e.g., $150 - $250 max, Under $1000]
* Essential Features (Non-negotiables): [LIST 2-3 MUST-HAVE FEATURES, e.g., Must have mapping, Must connect to Bluetooth, Must have lumbar support]
* Nice-to-Have Features (Bonuses): [e.g., Self-emptying bin, long battery life, aesthetic design]
* Priority (Quality vs. Price): [e.g., Prioritize maximum durability over lowest price, OR, Prioritize best value under $200]
Please deliver a complete Curated Buying Guide structured into the following 4 distinct sections:
1. THE TOP 3 RECOMMENDATIONS (The Shortlist)
Present 3 product recommendations that fit my criteria, following this proven comparison framework:
* Option A (The Best Overall / Editor’s Choice): The highest-quality option within budget.
* Option B (The Best Value / Bang-for-the-Buck): The best balance of features and price.
* Option C (The Budget Pick / Reliable Alternative): The most capable low-cost option.
For each option, include:
* Product Name & Approximate Price.
* 2 Key Pros & 2 Critical Cons (compiled from trusted expert reviews).
* Who this product is BEST for.
2. SPECIFICATION DEEP-DIVE TABLE
Construct a clean Markdown table comparing the 3 shortlisted options across these 5 columns: | Feature | Option A | Option B | Option C | Why It Matters |
3. BUYING CRITERIA & WHAT TO AVOID
* Define the 2 most critical "make or break" specifications for this product category that I must check (e.g., 'Suction Power (Pa)' for vacuums, 'Driver Size' for headphones).
* Identify 1 common marketing gimmick or trap in this category that I should ignore.
4. REAL-WORLD REVIEW CONSENSUS (The "Vibe Check")
* Summarize the common sentiment found in recent verified customer reviews regarding the reliability and typical failure points of these recommendations.
TONE & EXECUTION GUIDELINES:
* Approach this with an unbiased, analytical, consumer-advocate tone.
* Use clear bullet points and Markdown tables to make the recommendations easy to scan and act upon.Is this your AI Workout of the Week (WoW)? Cast your vote!
That's all we've got for you today.
Did you like today's content? We'd love to hear from you! Please share your thoughts on our content below👇
What'd you think of today's email?
Your feedback means a lot to us and helps improve the quality of our newsletter.
