Sponsored by
Moda is the AI design agent with taste
Moda's viral launch hit 4.4 million views in two days. Tens of thousands of professionals signed up. Startups, agencies, forward-thinking brands and top firms are now using Moda to create brand-aligned slides, ad creative, reports, social carousels and more.
Most AI tools tend to create what we call "AI slop": repetitions of the same colors, layouts and fonts. And when you try to fix it, you get stuck in a loop of re-prompting.
Moda is different. Drop in your website URL, and Moda learns your brand from the ground up: your colors, your fonts, your visual language. Then it helps you generate pro-quality slides, docs, and marketing assets.
The best part? Every layer is fully editable on a real canvas, and exports to powerpoint, PDF and more.
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
NASA Built an AI Doctor for Space, and It Got 88% of Diagnoses Right
TLDR: NASA is testing an AI helper that can diagnose and treat sick astronauts when they're too far from Earth to call a doctor.
The Story:
NASA is trying out an AI tool that acts like a backup doctor in space. It's called the Crew Medical Officer Digital Assistant, or CMO-DA, and it was built by NASA and Google. The whole point is to help crews who can't reach Earth in time. A message from Mars can take 20 minutes to reach Earth, and another 20 to come back. That 40-minute wait is a big problem if someone's hurt. So the AI runs fully offline, right on the equipment already onboard, no internet needed. In early tests, doctors rated its diagnoses at 88% accurate for ankle injuries, 80% for ear pain, and 74% for flank pain. It can take in symptoms, photos, and test results, then suggest what might be wrong and what to do next.
Its Significance:
Earlier this year, the agency brought a crew back from the space station early because of a medical worry. On a trip to the Moon or Mars, coming home early won't be an option. Right now astronauts on the space station can talk to a doctor on the ground almost instantly. Once crews head deeper into space, that assistance is gone. An AI that can help diagnose problems on the spot could be the difference between a small scare and a real emergency. The same idea could help people on Earth too, like folks in faraway places where the nearest doctor is hours away. It's still early, and the AI is meant to back up a trained crew member, not replace one. But it shows how AI is moving into jobs where waiting for a human isn't always possible.
QUICK TAKES
The story: Paul Meade, the Apple vice president who ran the Vision Pro headset and led work on Apple's upcoming smart glasses, is leaving to join OpenAI's hardware team. He spent over 15 years at Apple and will now help build OpenAI's new line of AI devices.
Your takeaway: This is part of a bigger pattern. AI companies like OpenAI and Anthropic keep pulling top people away from older tech giants like Apple and Google. OpenAI already hired Apple's former design chief Jony Ive and several others. When the new AI players can offer more money and a chance to build the next big thing, the legacy companies have a hard time holding onto their best talent.
The story: Google put out a white paper asking for a "middle way" on AI regulation, somewhere between heavy rules and no rules at all. Google president Kent Walker said the debate is stuck on a "false choice." This comes after other AI bosses called for rules, then pushed back when the rules might hurt their own business.
Your takeaway: Here's the part that's easy to miss. Big companies like Google can afford huge legal teams to deal with whatever rules come. A small startup can't. So in a world with heavy AI regulation, many small companies end up with two choices: get bought by a giant or shut down. Rules meant to keep AI safe can also quietly hand more power to the companies that are already on top.
The story: Chinese security company Qihoo 360 says it made an AI tool that finds software flaws better than Anthropic's Mythos model. The company's CEO called Mythos a "cyber nuclear weapon" because the US bans foreign access to it. He says China needs its own version as a deterrent, and that his team used a "swarm" of AI agents instead of one giant model.
Your takeaway: Nobody outside the company has tested these claims, so take them with a grain of salt. But the framing matters. A tool that finds security holes in software can be used to defend systems or to attack them. When two countries both race to build that kind of tool, it starts to look a lot like an arms race, just with code instead of missiles.
TOOLS ON OUR RADAR
🤖 Droidrun Free and Open Source: A groundbreaking mobile automation framework that enables you to control Android and iOS devices using natural language commands to perform complex repetitive tasks or data collection.
🔍 KnowTree Freemium: A unique conversation mapping tool that transforms your flat chat history into branching trees allowing you to visually track and compare different paths of thought.
📈 AtTheRate.ai Paid: A self learning marketing platform that monitors your campaigns across all major advertising channels to detect wasted spend and provide autonomous budget optimizations for better returns.
🎥 Knowlify Freemium: A generative video platform that converts your static documents and presentations into professional narrated explainer videos with automatically matched visuals and perfect pacing.
TRENDING
AWS CEO says replacing young workers with AI is "one of the dumbest ideas" — Amazon Web Services boss Matt Garman says cutting junior staff for AI is bad business: they're the cheapest workers, they know AI tools best, and they're your future senior staff. Honestly, this should be common sense. If you stop training new people, you run out of experienced ones later.
AI helped the FBI investigate the White House Correspondents' Dinner attack — A forensics firm says the FBI used its AI platform in the 48 hours after the attack to dig through seized phones and digital trails. Investigators can ask it plain questions like "show me all images where this suspect appears."
Google limited Meta's use of its Gemini AI — Google told Meta it couldn't sell all the computing power Meta wanted, which slowed some of Meta's AI projects. Meta even told staff to use AI "tokens" more carefully. Turns out even the giants are running short.
AI demand is outstripping supply, and even Google can't keep up — Big Tech is spending over $700 billion this year on AI infrastructure, yet it still can't build fast enough. Google Cloud is sitting on a $460 billion backlog of orders. The problem isn't that people don't want AI. It's that there isn't enough computing power to go around.
AI read a 2,000-year-old scroll burned by Mount Vesuvius — A scroll charred into a lump of charcoal in 79 AD has been read start to finish without ever being unrolled. Researchers used CT scans and machine learning to spot the ink, revealing a Greek philosophy text. Opening it by hand would've turned it to dust.
A mom on why AI is a "welcome guest" in her home — She writes that her kids now see their parents thinking out loud with AI instead of doomscrolling. Her 13-year-old uses Claude as a math tutor, builds Roblox games with it, and is sharp enough to spot fake AI videos by the bad spelling in captions. Her point: the tool reflects the judgment of the person using it.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
✉️ A live inbox dashboard that shows who fills your inbox most, who you owe replies to, and where your attention is really going.
Build a single-file HTML dashboard (vanilla HTML/CSS/JS) called Sender Leaderboard that connects to the user's real Gmail and visualizes where their inbox attention goes. Pull LIVE data via the Anthropic API with the Gmail MCP connector (no static mock data in the real build).
Data fetch: on load, POST to https://api.anthropic.com/v1/messages with model 'claude-sonnet-4-20250514', max_tokens 2000, and mcp_servers: [{ "type":"url", "url":"https://gmailmcp.googleapis.com/mcp/v1", "name":"gmail" }]. In the user message, instruct the model to analyze roughly the last 30 days of the inbox and return ONLY raw JSON (no preamble, no markdown) with this shape: { stats:{received, sent, unread, awaiting_reply, received_delta_pct, sent_delta_pct}, top_senders:[{name,email,count,type:"person|app|list"}] (top 10 by volume), reply_debt:[{name,subject,days_waiting}] (threads awaiting the user's reply, oldest first), categories:[{label,pct,color}] (Work/Notifications/Newsletters/Personal/Promotions summing ~100), volume_7day:[{day,n,today}], balance:[{name,recv,sent}] (top contacts, sent vs received). Parse the response by filtering data.content for blocks where type==='mcp_tool_result' to get the raw Gmail data, and read the model's final type==='text' block for the JSON; strip ```json fences and JSON.parse it. Wrap in try/catch with a clear error + "reconnect Gmail" hint on failure, and show a loading state while fetching.
Aesthetic: modern analytics dashboard. Near-black (#0a0d14), indigo→violet accent (#6366f1 / #8b5cf6) with a top-right radial glow, JetBrains Mono for all numbers/labels, Inter for text. Rounded card grid, gradient brand mark.
Render: topbar (brand mark + title + a "Last 30 days" range pill), a 4-stat row (Received / Sent / Unread / Awaiting your reply, each with a colored up/down/flat delta), then a two-column grid: left = Top Senders leaderboard (rank, name with person/app/list tag, email, gradient volume bar, count); right = "You Owe Replies" list (name + subject + age pill colored hot/warm/cool by days) and a category donut (SVG stroke-dasharray ring with total in the center) + legend. Below: a 7-day daily-volume bar chart (today highlighted) and a "Give & Take" panel showing each top contact as a split sent/received bar with a verdict (You owe / Balanced / You drive it). All charts hand-built with inline SVG and divs, no external libraries. Requires the Gmail connector to be enabled.What this does: Connects to your Gmail and analyzes the last 30 days to surface where your inbox attention actually goes. A top-stat row shows received, sent, unread, and reply-debt counts with week-over-week deltas. The centerpiece is a ranked Top Senders leaderboard with volume bars and person/app/list tags, alongside a "You Owe Replies" list sorted by age, a category donut breaking down work vs notifications vs newsletters, a seven-day volume chart, and a "Give & Take" panel showing whether each top contact is someone you drive the conversation with or owe. It pulls fresh data each time it loads via the Anthropic API and the Gmail connector.
What this looks like:
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Diagnose common medical problems from symptoms and photos with up to 88% accuracy, fully offline, for crews with no doctor nearby.
❌ Still Can't: Replace a trained human doctor. It's built to back one up, and accuracy still varies a lot by the type of injury.
✅ AI Can Now: Read text off a 2,000-year-old scroll too fragile to open, just from a CT scan.
❌ Still Can't: Reliably tell you the right specs for an e-bike. As one 13-year-old found out, it'll confidently make stuff up.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
Aurora by Kim Stanley Robinson - Book
A 2015 hard sci-fi novel about a generation ship that has been hurtling toward the Tau Ceti system for almost 200 years and is finally beginning its deceleration, with the chief engineer's daughter coming of age aboard as the ship's biomes start to break down faster than the crew can fix them. The most striking thing is the narrator: Robinson hands the entire book over to the ship's AI, which has to figure out how to write a novel about the humans it's responsible for.
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.
What if AI found your next job overnight?
Stop wasting hours filling out repetitive applications. AIApply automatically discovers relevant jobs, optimizes your resume, generates personalized applications, and applies for you around the clock so you can focus on securing your next opportunity faster.