In partnership with
The LA Mayor Market Ends June 2. The Odds Are Moving Now.
Karen Bass is at 68%. Spencer Pratt at 27%. Over $21 million is already trading on the Los Angeles Mayor race on Kalshi — and the prices are shifting every day as local news breaks. The window to get in before the consensus hardens is closing. The best prediction market trades happen before everyone agrees on the outcome. That moment is right now, before June 2. Peer-to-peer, no house, cash out anytime. Trade $10, get $10 free to start.
Trade responsibly.
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
OpenAI's New Biodefense AI Is Free, and It's Built to Catch the Next Pandemic
TLDR: OpenAI is giving governments and approved researchers free use of its strongest science AI to help spot and stop the next pandemic.
The Story:
On May 29, OpenAI launched a program called Rosalind Biodefense. It hands out free access to GPT-Rosalind, a model built to reason about molecules, proteins, genes, and disease. OpenAI says it'll pay all the costs, so price isn't a barrier. The program has two tracks: one for approved developers building early-warning and screening tools, and one for U.S. government agencies and allies working on outbreak response and new treatments. Early partners include Lawrence Livermore National Laboratory, Johns Hopkins, and the vaccine group CEPI, which plans to use it on threats like a rare strain of Ebola. OpenAI says GPT-Rosalind beats its own top models (GPT-5, GPT-5.2, and GPT-5.4) on science tests, and it briefed the White House before going public.
Its Significance:
The same kind of AI that could help someone build a bioweapon could also help scientists stop one. OpenAI calls its plan "defensive acceleration," which means putting the best tools in the hands of the people who protect us, before anyone else gets them. If it works, teams could design vaccines and catch outbreaks faster, and that affects your health the next time a new virus spreads. One worry stays open: a single company now helps decide who gets this technology and who doesn't.
QUICK TAKES
The story: Researchers at Washington University built an AI tool that reads a single blood draw and separates Alzheimer's, Parkinson's, frontotemporal dementia, and Lewy body dementia with over 90% accuracy. It can even spot when someone has more than one at the same time, which is common and hard for doctors to catch.
Your takeaway: Many people with memory problems never get the right diagnosis. A simple blood test that sorts this out could mean earlier, better treatment for millions.
The story: Google's new AI agent, Gemini Spark, runs in the cloud all day and can act on its own across Gmail, Calendar, Drive, and Docs, doing things like pulling deadlines from your inbox or writing a report from your notes. It runs on Google's Gemini 3.5 model and is opening up to people on Google's $100-a-month AI Ultra plan in the US.
Your takeaway: This is the move from chatbots that answer to agents that act for you. Handy, but you're handing an AI the keys to your email and files, so you trade some control for the convenience.
The story: John Koetsier looks at home robots like the 1X Neo that can tidy up, but they aren't fully self-driving yet, so a human worker often steers them from afar through the robot's cameras. The trade on the table: cheaper or free help in exchange for letting the company see inside your home to train the AI.
Your takeaway: A robot maid sounds great until you realize a stranger may be watching through its eyes. Free help isn't free if you're paying with your privacy.
TOOLS ON OUR RADAR
✨ KStars Free and Open Source: A digital planetarium that accurately simulates the night sky allowing you to explore millions of stars planets and distant galaxies from the comfort of your computer screen.
🎬 Synfig Free and Open Source: A two dimensional animation studio that uses vector graphics to help you create cinematic quality cartoons without needing to draw every single frame individually.
🎵 Rosegarden Free and Open Source: A beautifully designed digital audio workstation and musical composition environment that helps you write sheet music and arrange complex electronic melodies.
🔒 VeraCrypt Free and Open Source: A security application that creates completely encrypted digital vaults on your computer to keep your most sensitive personal files safe from unauthorized access.
TRENDING
Coders are refusing to work without AI, and that could come back to bite them - A research lab called METR couldn't even rerun a study because developers wouldn't code without AI anymore, not even for a test. But earlier work found AI made coders feel faster while actually slowing them down, since they spent extra time fixing its mistakes.
Cognition's Scott Wu says AI coding agents shouldn't replace humans - The CEO behind the Devin coding agent says his tool works at about a junior engineer's level and was never meant to take coders' jobs. In a year full of AI layoffs, he calls Devin a "buddy who helps you build more."
AI is fueling a new obsession with the afterlife - More companies are using AI to recreate the voices and faces of people who've died, letting families "talk" to lost loved ones. Some find it comforting. Critics worry these "deadbots" blur what's real and could be sold for profit.
MiniMax plans China IPO as it eyes local rivals like DeepSeek - Chinese AI startup MiniMax is preparing to sell shares on a mainland China market, months after its Hong Kong listing, to raise money for building bigger models. It's part of a wave of Chinese AI firms going public to keep up with rivals like DeepSeek.
Clinical chemistry analyzer market expands as demand for early disease detection and lab automation surges - The machines that test your blood and urine in labs are a growing business, projected to pass $20 billion by 2034. AI and automation are a big reason, helping labs run more tests, cut errors, and catch diseases earlier.
What is an AI prompt injection attack? - This guide breaks down a sneaky hack where someone hides secret instructions in text, files, or web pages so an AI follows them instead of you. It's now ranked the top security risk for AI apps, and there's still no perfect fix.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
✒️ Describe a decision you're about to make. Log your reasoning and predictions so future-you can audit it later.
Build a single-file HTML app with vanilla HTML/CSS/JS. The Decision Journal — log reasoning + predictions for future audit. Persist to localStorage key 'decision_journal_v1'.
Aesthetic: warm paper cream (#f5f3ee), subtle ruled-paper lines, Newsreader serif for body, Inter for UI, JetBrains Mono for labels. Burnt orange (#b85028) accent. Editorial journal feel with § section markers. Two-tab interface: New Entry / The Journal. Yellow-tinted hidden assumption card. Dark "for future-you" audit card with bright orange accents.
Form: decision textarea, type dropdown, reversibility dropdown (fully → one-way door), reasoning textarea (with help text "not the polished version"), alternatives textarea, predictions textarea (3 specific predictions), confidence slider 10-95 with live %, emotion input.
System instructions to the model: act as a decision journal assistant. Don't make the decision for them — help them log it cleanly for future-self audit. Honest, not flattering. If reasoning is weak or predictions vague, expose that. Return raw JSON: title (clean 1-line concrete), reasoning_summary (2-3 sentences distilling their reasoning), predictions array (3 items: restate as testable claim + confidence_label high/medium/low), watch_fors array (3 specific observable things that suggest going wrong), hidden_assumption (1-2 sentences naming the strongest unspoken assumption — honest not gentle), audit_opener (2-3 sentences to future-self, wise friend not judge tone), audit_questions array (4 specific questions to ask in 6 months).
Render: numbered entry header with date and confidence summary. White reasoning card. Orange-tinted predictions block with HIGH/MED/LOW confidence pills. Watch-fors card with ⚐ flag bullets. Yellow-tinted hidden assumption card with ⚐ marker. Dark audit card with audit_opener paragraph + bulleted audit questions in italic. Journal view shows all past entries as clickable cards with entry number, title, reasoning snippet, date, and confidence badge.What this does: Type the decision, your reasoning, the alternatives you rejected, three predictions about what will happen, your confidence level, and how you're feeling. Get back a numbered journal entry with your reasoning distilled, predictions made testable with high/medium/low confidence labels, three watch-fors (things that would prove you wrong), the hidden assumption underneath your logic, and four audit questions for future-you to ask in 6 months. Toggle to The Journal view to revisit past entries. Saves to localStorage.
What this looks like:
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Read one blood sample and tell four different brain diseases apart with over 90% accuracy.
❌ Still Can't: Run a home robot fully on its own, which is why companies still need humans steering through the cameras.
✅ AI Can Now: Reason about molecules and proteins well enough to help design vaccines and spot outbreaks faster.
❌ Still Can't: Reliably tell your real instructions apart from hidden malicious ones buried in a web page or file.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
Bong Joon-ho's follow-up to Parasite, about a man on a frontier space colony who keeps getting killed doing dangerous jobs and printed back into a new body each time, until version 17 and version 18 accidentally exist at the same time.
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.
Moda is the AI design agent with taste.
Your designer is busy. Agencies are expensive. And the deck needs to go out today.
Moda is your AI design partner for the moments when you need great work fast.
Drop in your brand, tell Moda what you need, and get polished, on-brand slides, docs, or ad creative in minutes.