Sponsored by

Welcome back, AI Admirers!

Breaking News: Apple just launched a research app that lets up to 200,000 US iPhone users ask an AI their health questions for more than 12 months. OpenEvidence powers the answers, and Apple hasn't said whether they will ever reach the Health app or Siri.

Get ready to dive into the latest happenings in AI.

📢 Today's Headline:

  • Apple's AI health study

  • Anthropic's IPO prices Wednesday

  • 6 robots run a store for 24 hours

  • Try This: Fact-check an AI health answer

  • Free Tools, Quick Hits & Money Moves

  • Today’s Poll and Results

Read time: 4 minutes.


Together With


AI agents are no longer just answering questions. They’re writing code, calling tools and touching production systems, and that changes what it takes to run them safely. Controls designed for human users and static apps weren’t built for software that decides what to do next.

Every new system an agent can reach is one more thing to secure. With that in mind, three trends are set to shape how enterprises run agents in 2027:

  • Secure sandboxes: agent-generated code runs in isolated environments, so a mistake or a bad output stays contained. It’s the first line of defense when agents run code you didn’t write.

  • Multi-agent workflows: protocols like MCP and A2A let specialised agents share a workflow without sharing every system, so each agent gets what it needs and nothing more.

  • Agent identity and access: managed identities and audit logging make every action traceable to the agent behind it, an area NIST and cloud providers are moving to address.

Each one comes down to the same idea: give agents room to work, with clear limits on what they can reach. Teams that build these in from day one can give agents more responsibility with less risk.


Hand-picked News


Apple wants 200,000 people to ask an AI health questions and grade every answer.

  • What happened: Apple launched a research app on Sunday that lets US iPhone users ask health questions and rate the answers for accuracy, detail and relevance. OpenEvidence powers the answers, which draw on journals like NEJM, JAMA and Nature.

  • Why it matters: Apple hasn't said whether these answers will reach the Health app or Siri, but this study is how it will find out. Users can optionally share HealthKit data, such as sleep, heart rate and lab results, for more tailored answers.

  • What’s next: The study runs for more than 12 months. It's open to US users aged 21 or older, on iOS 26 or later, with Apple Intelligence turned on.

Apple isn't selling an AI doctor yet. It's collecting the homework first.

💡 What it means for you: Apple says study data is anonymized and won't train AI models, and OpenEvidence deletes queries after 24 hours. Read those terms before you opt in any HealthKit data.


Orders are reportedly several times the shares on offer.

  • What happened: Bloomberg reported that orders for Anthropic's IPO are multiple times the 825M shares on offer, driven largely by big institutional orders. The range is $70 to $78 a share, with pricing set for Wednesday, Nov 18.

  • Why it matters: At $78, Anthropic would be worth about $1.02T, or about $1.4T fully diluted. Bloomberg notes that falls short of some investors' hopes, so the cautious range leaves room for a higher final price.

  • What’s next: Trading is set to start Nov 19 on the Nasdaq under the ticker "ANT."

Cautious range, hungry buyers. Wednesday shows which one wins.

👀 What to watch: Whether the final price lands above the $78 top of the range on Wednesday.


They finished 1,730 tasks in 24 hours. Two of the six still failed grading.

  • What happened: Six AgiBot G2 wheeled humanoids ran a live Shanghai convenience store for 24 hours, handling jobs from coffee orders to shelf restocking. They completed 1,730 tasks in total.

  • Why it matters: A live store is a tougher test than a stage demo. With two of the six failing grading, the robots look close to useful but not yet reliable.

  • What’s next: Demand is already building. JD.com reports humanoid robot sales grew 700% year over year during China's Double 11 shopping festival.

Robots can run a shop. Just not every shift, yet.

🧠 Our take: Real stores are the new benchmark, and a 4-out-of-6 pass rate says the hype is still ahead of the hardware.


Free Tools & Resources

🎓 Elements of AI Free

A free online course that explains what AI is, what it can and can't do, and how it touches everyday life.

Best for: beginners who want the basics without the jargon.

Courses on building AI agents with LangChain. The site says the courses are always free, with no account needed.

Best for: developers who want to build their first agent.

💻 Ollama Free

Run AI models on your own computer. Local use is free and unlimited, so your questions never leave your machine. Its cloud models are pay-per-token.

Best for: privacy-minded users who want to keep their data at home.

A hands-on course on how large language models work, using Hugging Face tools. The page says it is completely free and without ads.

Best for: readers who already know Python and want to go deeper.


Quick Hits


Try This Today

⚡ Fact-check any AI health answer in 2 minutes

Apple is testing AI health answers at huge scale, so here is how to test one yourself. Copy this prompt into ChatGPT, Claude or Gemini:

❝

Here is a health answer an AI gave me: [paste the answer]. Act as a careful reviewer. List each claim and say whether it is well supported, uncertain or likely wrong. Name the kind of source that would confirm it. Then give me three questions to ask my doctor. Do not diagnose anything.

Works for anyone who asks AI about health, fitness or nutrition. It is a reality check, not medical advice.


Quick Poll

Today’s Poll (one tap to vote):

Apple is testing an AI that answers health questions, and can read your HealthKit data if you opt in. Would you join the study?

Login or Subscribe to participate

Vote today, see the results tomorrow!

Previous Poll:

ChatGPT Wallet lets an AI agent pay for things after a Face ID check. Would you let it shop for you?

  • A: Yes, with a spending cap 💳 – 14%

  • B: No, I'll click buy myself 🙅 – 86% 🏆

❝

Only 7 of you voted, but 6 said you want to press the buy button yourself. Trust, not technology, is what holds back AI payments right now.

The Simplified AI team


By the Numbers

👥 200,000

The most people Apple will let into its AI health study, which runs for more than 12 months. Startup Fortune

💰 $90.4B

The total Anthropic could raise including selling shareholders, which would top SpaceX's record June IPO. StockAnalysis

🤖 1,730

The tasks six humanoid robots finished while running a Shanghai convenience store for 24 hours. The Robotics Media


Money Moves

💵 Anthropic: $26B private placement - Seven companies, including Alphabet, Nvidia, Microsoft and TSMC, buy in at the IPO price.

🏭 TSMC plans a $500M fund - It would back chip startups and upstream suppliers, according to Nikkei Asia.


From our Partner

A CRM so smart, it updates itself.

HubSpot's CRM is so smart, it now updates itself. Calls get logged and summarized the moment they end. New leads get researched and a first outreach drafted, automatically. Deals move forward on their own, with next steps flagged before you have to ask. 

Even the one repetitive task you've been meaning to fix can now run itself, no code required. Customers using it are already seeing the difference. 

This isn't the CRM you have to work hard to figure out. It's the one already working before you log in.


Only AI You Need

AI is now answering health questions and running shops. It can also fix the oldest item on your to-do list: a better profile photo. HeadshotPro turns a few selfies into professional headshots, usually within an hour.

📌 Key Features

  • Easy: upload a few selfies, no studio needed.

  • Variety: get 30 to 70 professional headshots.

  • Fast: results usually arrive within an hour.

  • Price: from $29, one-time.

  • Teams: plans for companies that need matching photos.

Offer: money-back guarantee, and HeadshotPro says 250,000+ customers have used it.

Affiliate link: we may earn a commission at no extra cost to you.


💬 One quick favor

Hit reply and tell us: Would you let an AI read your health data for a study, and what would it take to say yes? We read every reply, and it helps your inbox know you want these emails.

Know someone who’d like this? Forward it to them. See you tomorrow!