Sponsored by
How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads
The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.
Beginners in AI
Good morning and thank you for joining us again!
Welcome to this daily edition of Beginners in AI, where we explore the latest trends, tools, and news in the world of AI and the tech that surrounds it. Like all editions, this is human curated and edited, and published with the intention of making AI news and technology more accessible to everyone.
THE FRONT PAGE
AI Outperformed ER Doctors at Harvard, But Your Doctor Isn't Going Anywhere
TLDR: A new Harvard study found that OpenAI's o1 model gave the right diagnosis 67% of the time on real ER cases, beating two human doctors who scored 55% and 50%.
The Story:
The study came out this week in the journal Science. Researchers at Harvard Medical School and Beth Israel Deaconess Medical Center looked at 76 real patients who walked into the Beth Israel emergency room. They had two regular doctors write diagnoses for each patient. They also had OpenAI's o1 and 4o models do the same thing using the exact same patient records.
Then two other doctors graded the answers without knowing which ones came from a person and which came from AI. The o1 model got the right or close-to-right diagnosis 67% of the time. One human doctor hit 55%. The other hit 50%. The gap was biggest at first triage, when the AI had the least info to work with. In one case, the AI flagged a possible history of lupus that the human team missed, which helped explain why a patient kept getting worse.
Its Significance:
Headlines like this can make people panic that doctors are about to lose their jobs. They're not. Look at radiologists. People predicted AI would replace them years ago, and there are now more radiologists working than before. The same thing is likely here. AI will help doctors catch things they miss, especially in busy ERs where mistakes happen because people are tired and rushed.
The study authors said their own work doesn't show AI is ready to make life-or-death calls alone. It shows AI could be a strong second opinion. For patients, that means fewer missed diagnoses and faster answers. For doctors, it means a tool that handles the boring pattern-matching so they can focus on the parts of the job that need a real human, like talking to scared patients and making judgment calls. This is how it should be. AI to enhance rather than replace.
QUICK TAKES
The story: A new report from the Foundation for Defense of Democracies says China is building groups of AI-powered four-legged robots that work together like a wolf pack. They scout, carry supplies, and may fight alongside soldiers. The report names Taiwan as the most likely place China would use them.
Your takeaway: Robot warfare isn't science fiction anymore. These machines have weak spots though. They run on batteries and need radio links, so jamming and cyberattacks could shut them down. The unanswered question for everyone: when a robot makes a battlefield mistake, who's responsible?
The story: Stanford bioengineering professor Emma Lundberg is leading a global push to build the first AI-powered virtual human cell. The model would simulate how real cells behave, which could help scientists test new drugs without needing animal or human trials first.
Your takeaway: Lundberg calls this the holy grail of biology and admits a working version could take ten years or more. But if it works, drug discovery could go from a 10-year process to something far faster, and people could get treatments tailored to their own cells.
The story: Meta bought Assured Robot Intelligence, a small startup with about 20 employees building AI brains for humanoid robots. The team joins Meta's Superintelligence Labs. Meta has already bumped its 2026 spending plan up to between $125 billion and $145 billion, with much of that going to physical AI.
Your takeaway: Meta wants to be the Android of robots. The plan is to build software that other robot makers license, not just sell one branded robot. That puts Meta in a direct race with Tesla's Optimus and Amazon, which bought a similar startup called Fauna last month.
TOOLS ON OUR RADAR
🖨️ Cura Free and Open Source: A remarkably powerful slicing software used by hobbyists around the world to prepare their three dimensional digital models for actual physical production on a desktop 3D printer.
💡 LightBulb Free and Open Source: A clever application that protects your eyes by automatically adjusting the color temperature of your computer screen as the sun sets in your local time zone.
🧸 GCompris Free and Open Source: A wonderful educational software suite packed with hundreds of fun activities and mini games designed specifically to help young children learn math geography and computer skills.
🎙️ Mumble Free and Open Source: A private and low latency voice chat application designed for gamers and large groups of friends to talk to each other clearly over the internet.
TRENDING
Chinese Court Says Companies Can't Fire Workers Just to Replace Them With AI — A Hangzhou court ruled that swapping a worker for a large language model isn't a legal reason to end a job contract. The case involved a quality assurance supervisor named Zhou who refused a 40% pay cut and won twice in court.
New York City Cancels AI High School After Parent Backlash — The planned Next Generation Technology High School was pulled before its vote. Parents and teachers worried about putting kids in an unproven AI-first classroom. One advocate called it the biggest education uproar she's seen in 25 years.
"This Is Fine" Cartoonist Says AI Startup Stole His Art — KC Green, who drew the famous burning-room dog meme back in 2013, says AI startup Artisan put his art on a subway ad without asking. The same company has run billboards telling businesses to "stop hiring humans." Green is now looking into legal help.
Western States Are Adding AI Cameras to Spot Wildfires Faster — Arizona Public Service has nearly 40 active AI smoke-detection cameras, with 71 planned by summer. Pano AI cameras spotted 725 wildfires in the U.S. last year and warn crews about 45 minutes faster on average than the first 911 call.
Kennesaw State Student Builds AI Tool to Speed Up Breast Cancer Detection — Preston Brantley, a double-major in mechanical engineering and nursing, is training an AI to scan microscope slides and flag possible cancer cells. The goal isn't to replace pathologists. It's to cut down a wait that can stretch from weeks to months.
Almost 40% of New Podcasts Last Week Were Probably AI-Generated — Bloomberg, citing the Podcast Index, reports that 39% of new podcast feeds over a recent nine-day stretch were likely made by AI. One company, Inception Point AI, is putting out around 3,000 episodes a week across 10,000 shows. Critics call it "podslop."
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
⚡ Describe yourself badly. Get rewritten as a legend, a myth, and a force of nature.
Build a single-file HTML app using vanilla HTML, CSS, JS, and one API fetch call. Create The Hype Machine — a tool that rewrites someone's modest self-description as a legend. Use localStorage key 'hype_machine_v1'. No branding from any AI company anywhere in the UI.
Aesthetic: near-black (#0a0a0a) with a subtle yellow grid overlay, scanline texture, Anton condensed display font for headings, Space Grotesk for body, Space Mono for labels. Gold (#ffd700) accent throughout. The main title has a 4px yellow drop shadow. A pulsing gold badge in the header.
Form: large textarea for the bad self-description, fields for craft/work and one real win, a mode dropdown (All Three / Legend / Myth / Force).
Call the API with a system prompt that instructs it to reframe — not invent — finding genuine power in what the person dismisses as ordinary. Return raw JSON: one_line (the machine's core read in 10-15 words), legend (bombastic declarative, short punchy sentences, all-caps energy), myth (poetic, ancient, metaphors from nature and time), force (second person present tense, specific and urgent).
Render: a comparison strip showing their original vs the machine's one-line read, three tab panels (Legend / Myth / Force) each with distinct typography (Anton all-caps for Legend, italic light for Myth, bold present-tense for Force), Copy Active Mode and Copy All Three buttons, a Hall of Fame section below from localStorage. Make it work in a single HTML file.What this does:
Type your most underwhelming self-description — the couch, the half-finished projects, the late deadlines. The Hype Machine finds what is actually there and rewrites it three ways: The Legend (bombastic, declarative, all-caps energy), The Myth (poetic and ancient, as if you were written about before you were born), and The Force (second-person, present-tense, specific and urgent). A comparison strip shows what you said versus what the machine sees in one line. Every activation saves to a Hall of Fame in localStorage so you can return to your best version on the days you forget it.
What this looks like:
Payroll errors cost more than you think
While many businesses are solving problems at lightspeed, their payroll systems seem to stay stuck in the past. Deel's free Payroll Toolkit shows you what's actually changing in payroll this year, which problems hit first, and how to fix them before they cost you. Because new compliance rules, AI automation, and multi-country remote teams are all colliding at once.
Check out the free Deel Payroll Toolkit today and get a step-by-step roadmap to modernize operations, reduce manual work, and build a payroll strategy that scales with confidence.
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Match or beat human ER doctors at first-look diagnosis on real patient cases when given the same medical records.
❌ Still Can't: Run an emergency room on its own. The Harvard authors said real-world trials are needed before AI makes any life-or-death calls without a doctor in the loop.
✅ AI Can Now: Spot the early signs of a wildfire from a camera feed about 45 minutes before the first 911 call comes in.
❌ Still Can't: Decide what to do once a fire is found. Human dispatchers still have to choose whether to send crews, monitor, or evacuate.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
A man wakes up trapped in an automated prison run by a polite AI warden named Howard, and the entire film becomes a quiet two-hander between the prisoner trying to escape and the machine trying to keep him alive. Writer-director Travis Milloy made it on a shoestring with essentially one actor and a voice, but it earns every one of its big questions about consciousness, simulation, and whether the AI across from you might actually be the only thing on your side. One of the smartest AI films of the last decade and almost nobody has seen it.
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.
