Sponsored by
You paid $5,000 for that website. You can't even update it
Agencies charge thousands. Take weeks. Hand you something that needs a developer every time you want to make a change.
Readdy builds you a professional, mobile-ready website in minutes, with SEO, hosting, booking, and payment integrations included. You just need to describe your business, and when you need to update something, you just need to ask our AI. No developer call. No extra invoice.
You get the same polished result at a fraction of the price. And it’s all done before your agency would have sent the first draft.
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
A New AI Tool Lets Chemists Build Molecules Just by Talking to It
TLDR: A new AI system called Synthegy lets chemists design new molecules just by describing them in plain English. Built by researchers in Switzerland, it plans the steps to make a molecule and explains its reasoning along the way. This could speed up the discovery of new medicines, plastics, and clean fuels.
The Story
Scientists at a Swiss research school built a new AI system called Synthegy that helps chemists make new molecules by typing what they want in plain English. Building a molecule used to take years of training and hundreds of small choices. Now a chemist can describe the molecule, and the AI maps out a step-by-step plan to make it. The system also explains why each step works, instead of just spitting out an answer. It scores different paths and picks the smartest one. The research came out in the journal Matter on May 5, 2026. You can read about it here.
Its Significance
This is underrated news because making new medicines, plastics, and clean fuels all start with making new molecules. If a chemist can talk to an AI like a partner instead of running a thousand tests by hand, new drugs and materials could show up much faster. The AI doesn't just guess. It reasons through the chemistry the way a senior scientist might. That moves AI from being a tool that fills in answers to being a thinking helper that works alongside experts.
QUICK TAKES
Bumble Is Killing the Swipe and Going All-In on AI
Bumble's CEO Whitney Wolfe Herd told Axios the dating app will get rid of swiping by the end of 2026. Paid users dropped 21% in the last quarter, and the company is betting on an AI dating helper called Bee instead. The plan is to use AI for matching people based on deeper info, not just photos. Some folks like the idea. Others worry about a future where your AI bot dates other AI bots for you. The full story is here.
Tether's New Medical AI Runs on Your Phone and Beats Models 16 Times Bigger
Tether, the stablecoin company, just dropped a medical AI called QVAC MedPsy that fits on a smartphone with no internet needed. The smaller version has 1.7 billion parameters and beat Google's MedGemma-27B, a model 16 times its size, on a tough clinical test graded by 262 real doctors. Because it runs on the phone, patient data never leaves the device. That makes it useful in rural clinics, field medicine, and any place where privacy or signal is shaky. See the story here.
Scientists Made AI Play Battleship to Help It Do Better Science
Researchers tested AI models on the board game Battleship to see how well they handle smart guessing with limited info. The trick: science also runs on smart guesses, since experiments cost money and time. One small model called Llama-4-Scout, helped by other AI partners, beat GPT-5 in fewer moves about two-thirds of the time and at one-hundredth the cost. Even better, the AI got smarter when the models talked to each other in code instead of plain English. Read the full piece here.
TOOLS ON OUR RADAR
📓 Anytype Free and Open Source: A beautiful private workspace that helps you build a digital second brain to organize your daily journals web clippings and creative ideas without relying on cloud servers.
🌙 Moonlight Free and Open Source: A phenomenal game streaming application that lets you play the high end video games installed on your main computer remotely from your smartphone or living room television.
🎧 Audiobookshelf Free and Open Source: A fantastic dedicated library server specifically designed to help you organize stream and track your progress through your massive collection of digital audiobooks and podcasts.
🎨 Tux Paint Free and Open Source: An award winning drawing program created specifically for children featuring a fun interface funny sound effects and an encouraging cartoon mascot to inspire creativity.
TRENDING
Mayo Clinic AI Spots Pancreatic Cancer Up to 3 Years Early
Mayo Clinic built an AI called REDMOD that catches signs of pancreatic cancer on regular CT scans up to three years before doctors usually find it. The team tested it on nearly 2,000 scans that radiologists had marked as normal. The AI found 73% of the early cancers, almost double what specialists caught on their own. Pancreatic cancer kills more than 85% of patients within five years, mostly because it gets caught too late. Read more here.
Chrome Quietly Deleted Its On-Device AI Privacy Promise
Google Chrome version 147 used to say its on-device AI ran "without sending your data to Google servers." That line is gone in version 148. The change comes after users found Chrome was silently downloading a 4GB file for the Gemini Nano AI model, even when they didn't ask for it. Delete the file and Chrome puts it back on the next restart. A privacy researcher says the silent install may break EU privacy law. Story here.
Your AI Chats Might Be Leaking to Meta, Google, and TikTok
Researchers at IMDEA Networks found more than 13 ad and tracking tools hidden inside ChatGPT, Claude, Grok, and Perplexity. Grok was the worst, with public chat links by default and TikTok's tracker grabbing actual message text. The study didn't prove anyone is reading your chats, but the wires are in place. Rejecting cookies sometimes helps and sometimes doesn't. Full report here.
OpenAI Rolls Out GPT-5.5-Cyber for Verified Defenders
OpenAI started letting trusted security teams use a new model called GPT-5.5-Cyber. It's a stripped-down version of GPT-5.5 with fewer guardrails, built for jobs like finding software bugs, reviewing patches, and reverse-engineering malware. Only vetted defenders get access through a program called Trusted Access for Cyber. The move follows Anthropic's similar Mythos release last month. Read the announcement here.
OpenAI Adds a Trusted Contact Feature for Self-Harm Alerts
ChatGPT now lets adult users pick a friend or family member to be alerted if the system spots signs of serious self-harm risk. The chosen person has to accept the role first, and the alert is short with no chat details shared. The rollout follows lawsuits claiming ChatGPT validated suicidal thoughts in vulnerable users, including a 16-year-old who died in 2025. See the story here.
Firefox Found Decade-Old Bugs Using Anthropic's Mythos
Mozilla just shared what happens when you point Anthropic's Mythos AI at Firefox's code. The model found a wealth of high-severity bugs, some hidden in the codebase for more than ten years. Firefox shipped 423 bug fixes in April 2026, compared to just 31 the year before. Mozilla still uses human engineers to write the actual patches because the AI's fix code isn't ready to ship as-is. Story here.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
⚗️ Describe a life change you're considering. Get a rigorous personal experiment with protocol, metrics, controls, and conclusion criteria.
Build a single-file HTML app using vanilla HTML, CSS, JS, and one API call. Create The Experiment Designer — a tool that designs rigorous n=1 personal experiments. Use localStorage key 'experiment_designer_v1'.
Aesthetic: warm off-white (#f0ede6) with horizontal ruled lines like a lab notebook, a faint red margin line on the left, IBM Plex Serif for body text and protocol copy, IBM Plex Sans for UI, IBM Plex Mono for labels and IDs. A stamped header badge. Clean scientific document feel. Box shadows on cards give a stacked paper effect.
Form with three sections (§ 1 Intervention, § 2 Baseline, § 3 Constraints): intervention input, why/hoped-outcome row, baseline description, health dropdown, duration dropdown (14/21/30/60/90 days), constraints and resources fields.
Call the API with a system prompt instructing it to be concrete and specific — no vague instructions, something a non-scientist can actually follow. Return raw JSON: experiment_id, title, hypothesis (proper if-then format), duration_days, frequency, difficulty, variables object (independent/dependent/control), protocol (3-4 paragraphs of exact instructions), schedule array (period + focus per phase), metrics array (name + exactly how to measure), controls paragraph, risks array, conclusion_criteria (specific measurable criteria for success/failure).
Render as a formal protocol document: dark header with experiment ID and hypothesis in italic, a four-cell stats row (days/frequency/metrics count/difficulty), variables in three cards, protocol as paragraphs, schedule as a two-column grid, metrics as numbered cards with measurement instructions, controls paragraph, risks as warning-prefixed list, conclusion in a left-bordered box. Archive saves all experiments. Make it work in a single HTML file.What this does:
Describe something you want to test on yourself — cold showers, sleep optimization, journaling, no alcohol, morning walks — and get back a proper scientific protocol. It generates a formal hypothesis, identifies independent and dependent variables, writes step-by-step instructions specific enough to follow without guessing, breaks the experiment into weekly phases, defines three measurable metrics with exactly how to track them, specifies what to keep constant to avoid confounds, flags potential risks, and gives you clear conclusion criteria so you know at the end whether it actually worked. Every protocol saves to localStorage with a lab-style experiment ID.
What this looks like:
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.
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Spot pancreatic cancer signs on CT scans up to three years before human radiologists catch them.
❌ Still Can't: Write security patches that ship straight to production. Even with a model finding hundreds of Firefox bugs, every fix is written and reviewed by a human engineer.
✅ AI Can Now: Take a description of a molecule in plain English and have and plan out the steps to make it, complete with reasoning for each one.
❌ Still Can't: Be assumed private. All four major chatbots quietly send some user data to ad and analytics trackers, even when cookies are rejected.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
A marine biologist studies an octopus species that may have developed its own language, while subplots follow an autonomous AI ship and the world's first true android. The book takes seriously that intelligence might come in forms completely alien to ours, and it asks whether we'd even recognize it when we found it. Won a 2024 Locus Award and still flies under the radar outside hardcore sci-fi circles.
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.
