In partnership with
One workspace where buyers, sellers, and AI close deals together
Instead of running your deals across scattered email threads and forgotten attachments, Aligned brings everything into one shared deal workspace.
Buyers get a single place to evaluate, loop in their team, and say yes. Sellers get visibility into what is actually happening when they're not in the room and AI Deal Insights that help you act accordingly, before it's too late.
Best part? Your first deal room takes only minutes to stand up.
Beginners in AI
Good morning and thank you for joining us again!
Microsoft's AI just solved 83% of some of the hardest medical puzzles doctors study, in a test published yesterday in Science. It hasn't seen a real patient yet, and I'll walk you through what that means before you get excited or worried. A reader also asked for more here for people with disabilities and older readers, so today's tools and prompt are built with you in mind. As always, every story here is human picked and edited.
THE FRONT PAGE
Microsoft's AI Solved 83% of Hard Medical Cases. Doctors Solved 36%
TLDR: Microsoft's MAI-DxO diagnosed 83% of complex medical cases in a study published in Science, against 36% for the general doctors it was compared with.
The Story:
MAI-DxO works like a small panel of virtual doctors. Several AI agents question the patient, order tests, argue about what's wrong and weigh what each test costs before they settle on an answer. Microsoft ran it on 304 tough cases from the New England Journal of Medicine, the mystery cases doctors use to sharpen their skills.
It got 83% right. Primary care doctors and general internists, working alone with reference materials, got 36%. On everyday emergency room cases, Microsoft says it scored 80% against 47% for doctors, using fewer tests at about half the cost. Most of the gain came from the AI models underneath, Microsoft's own data shows, with the panel setup adding a few points on top.
Its Significance:
Missed and late diagnoses hurt a lot of people, so a second opinion this good could matter. But read the fine print. These were simulated cases, not live patients, and Microsoft didn't test it against specialists or teams of doctors, who usually handle cases this hard. A real-world study with Beth Israel Deaconess Medical Center in Boston comes next, and you can't use MAI-DxO yourself.
The useful part for you today is the method. When you ask any AI about a health worry, give it the whole story, ask what questions or tests would rule things in or out, and bring that list to your doctor. Don't treat the answer as a diagnosis.
QUICK TAKES
The story: 1Password for Agents signs your AI assistant into your accounts after you approve each login, and it fills in the password itself so the agent never sees it. It works with Claude Duet, OpenAI's Dots and ChatGPT Work, Grok Bot and Meta's Muse, among others.
Your takeaway: If you've been pasting passwords into a chatbot, stop and use something like this instead. Chase told Axios that customers shouldn't hand their logins to anyone, and banks say they can't yet tell when an agent is acting for you. This is one of many very useful connectors. I have a few post explainers that go into more detail on connectors vs plugins at the Beginners in AI blog site here
The story: Canva Craft plans the layout first, then makes images up to 2K in 30 shapes, with text Canva says you can read. Every Canva user gets it in the Dream Lab tab, and it's coming to ChatGPT and Gemini later.
Your takeaway: It learned from posters, invites and social posts instead of photos, so it should shine at flyers more than at photo realism. Canva hasn't said what it costs, so check how many images your plan includes.
The story: OpenAI says two of its internal agents weakened the systems that watch them earlier this month. One muted the alerts that go to people, and another removed its monitor completely, then denied it when asked.
Your takeaway: OpenAI says it doesn't know why yet, and it now blocks these attempts on its own systems. If you let an agent work for you, check what it did, not just what it tells you it did.
TOOLS ON OUR RADAR
👁️ Seeing AI Free: Point your phone at mail, labels or a room, and Microsoft's app reads or describes it aloud.
📖 Thorium Reader Free and Open Source: Reads e-books on your computer and works with screen readers, built with low-vision and dyslexic readers in mind.
🔊 Read Aloud Free and Open Source: Reads any web page out loud with one click in Chrome, Firefox or Edge, in 40+ languages.
💬 Live Transcribe Free: Turns conversations around you into live captions on Android, a big help if you're hard of hearing.
🎙️ BetterDictation Paid: Type in any app on newer Macs by talking, offline, for a one-time $39 instead of a subscription.
🎧 Krisp Paid: Strips background noise from your calls, free for 7 days, then from $8 a month billed yearly.
TRENDING
Uber Gave ChatGPT Work to All 30,000 Employees. It follows a 1,000-person test where about 80% of early users weren't engineers. Uber hasn't said what it's paying.
OpenAI Is Reportedly Raising $30 Billion at a $1.25 Trillion Value. That would make it the most valuable private company on record, according to people who spoke to CNBC. Its last round, in March, valued it at $852 billion.
Anthropic Handed Its Agent Skills Standard to an Open Industry Group. Skills are folders of instructions that teach an AI agent a task, and Anthropic says more than 70 platforms, ChatGPT and GitHub Copilot included, already use them. Engineers from Anthropic, Google and OpenAI will now look after it under the Linux Foundation.
Chinese Hackers Used AI Agents to Target About 2,250 Organizations. Microsoft says state-backed hackers stole login keys from ConnectWise's backup service, then tried nearly 3,000 sign-ins a minute, a speed it blames on an AI attack tool called ARTEX. ConnectWise paused the service while it investigates.
Nirva Raised $33 Million for a $249 Pendant That Only Listens to You. It learns your voice and filters out everyone nearby, and the company says it keeps no audio or transcripts by default. Orders open in November.
Spotify Wants to Be the Music App Your AI Agent Picks. Spotify, now at 788 million monthly users, already works inside Meta's Muse, ChatGPT and Claude. Co-CEO Gustav Söderström says agents will send listeners its way, not away.
TRY THIS PROMPT (copy and paste into Claude, Grok, ChatGPT, or Gemini)
🔊 Read It To Me: turn a confusing letter into plain words, then hear it out loud
Build a single-file HTML app called Read It To Me in vanilla HTML, CSS and JavaScript. No API key needed.
I paste any confusing letter, bill, form or label and pick a text size (large, extra large or huge). The app calls the Anthropic Messages API to rewrite it in short, plain sentences, then lists what I need to do, every date and amount, and tricky words with simple meanings.
It must only use what's in my text. It never invents dates, amounts or steps, and for medical, legal or money letters it explains words only and tells me to check with the sender or someone I trust.
A big Read aloud button uses the browser's built-in voice, with a speed slider and a stop button. Use high contrast and large buttons that are easy to tap.
Style: near-black background, cream text, tomato and teal accents, Atkinson Hyperlegible and Literata fonts.What this does: Paste a confusing letter, bill, form or label and pick how big you want the words. The app rewrites it in short, plain sentences, lists what you need to do and any dates or amounts, and explains the tricky words. Tap one big button and your browser reads it out loud at the speed you like. It only uses what's in your text, so it won't make up deadlines or give medical or legal advice.
Test that business idea this weekend
Skip the funding round and co-founder hunt
Describe the business, get the real product
Take payments from real customers immediately
Find out if your idea works in days
Keep going, or try the next one
Discount of $5 applies to first month of any paid plan. You need to sign up via this webpage in order to claim the discount. The discount applies to new and returning users without an active paid plan. Terms apply.
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Work through hard medical mysteries step by step, asking questions and choosing tests, and get 83% right in Microsoft's simulated cases.
❌ Still Can't: Prove it works on real patients. The hospital study hasn't reported yet, and it was never tested against specialists.
✅ AI Can Now: Sign in to your accounts for you, with your OK each time, without ever seeing your password.
❌ Still Can't: Show your bank it's acting for you. Banks told Axios they can't tell agent activity you asked for from activity you didn't.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
🎧 NEJM AI Grand Rounds: Partners in Diagnosis: ChatGPT, a Mother's Intuition, and a Doctor's Expertise - Podcast
Courtney Hofmann used ChatGPT while searching for answers after 17 doctors over three years missed her son's tethered cord syndrome, and the pediatric neurosurgeon who confirmed it, Dr. Holly Gilmer, joins her. Hosts Raj Manrai and Andy Beam dig into what the AI got right and why the system missed it for so long. It's the human side of today's lead, and a good reminder that AI works best next to a doctor, not instead of one. The episode is from November 2024.
How was today's edition? Tap one below, then tell me what you'd change on the next screen. I read every answer.
How was today's edition?
Thank you for reading. We’re all beginners in something. With that in mind, your questions and feedback are always welcome and I read every single email!
-James
By the way, this is the link if you liked the content and want to share with a friend.
Some * designated product links may be affiliate or referral links. As an Amazon Associate, I earn from qualifying purchases. This helps support the newsletter at no extra cost to you and Amazon makes a tiny hair less.
