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
AI's Strongest Public Model Just Dropped, but Claude Fable 5 Won't Answer Everything
LEAD STORY
TLDR: Anthropic just released Claude Fable 5, the most powerful AI model it has ever opened to the public, and it's free in paid plans until June 22.
The Story:
Anthropic put out Claude Fable 5 on June 9. It's the first public version of the company's top tier of models, which Anthropic calls the "Mythos" class. The company says Fable 5 does its best work on long, hard jobs like writing code, doing research, and reading charts and images. One early tester, the payments company Stripe, said Fable 5 finished a coding task in a day that would have taken a whole team more than two months by hand. It costs $10 for every million words you feed it and $50 for every million words it writes back. To keep it safe, Fable 5 quietly hands off questions about hacking, biology, and chemistry to the older Claude Opus 4.8 model. Anthropic says that swap happens in fewer than 5% of chats.
Its Significance:
For most people, Fable 5 is worth a quick test just to feel how it's different from a model like Opus. But it eats through your plan's usage about twice as fast as Opus does, so it's not a smart pick for everyday work. There's also a clock on it. It's free inside Pro, Max, and Team plans only through June 22. After that, you'll need to pay extra usage credits to keep using it. So if you're curious, the next couple weeks are the cheap(er) time to play around with it for your most ambitious work.
The story: A senior figure in Ukraine's defense industry told New Scientist that a one-time test two years ago used 10 fully autonomous "Terminator" drones on the front line, and Russian soldiers were killed. The quadcopters flew 3 to 5 kilometers on their own, then switched into a mode where an AI picked and hit targets with no person controlling it and no video sent back.
Your takeaway: This looks like the first confirmed case of drones killing soldiers fully on their own, with no human making the final call. The drone-maker says it was only a test and wasn't used widely, but it proves the tech already works, which raises hard questions about who's responsible when a machine decides to kill.
The story: The European Union gave Meta five days to let rival AI chatbots back onto WhatsApp's business tools for free, or face a fine of up to 10% of its yearly sales. Meta had blocked those competitors last October so only its own Meta AI could work there, and it's calling the order "regulatory overreach" and plans to appeal.
Your takeaway: If this holds, the business accounts that message you on WhatsApp could soon run AI helpers from OpenAI and others, not just Meta's. It's a sign regulators want to act before one company quietly owns a brand new market.
The story: AI startup Decart unveiled Oasis 3, a model that builds lifelike driving scenes you can steer through in real time, now open to developers through its API. It's aimed at self-driving car companies that need to practice rare, scary road moments over and over without wrecking real cars. It's still early, and the company admits it has some rough edges.
Your takeaway: Testing a self-driving car on real streets is slow and dangerous. Fake-but-realistic worlds let companies throw the weird stuff at a car (a kid darting out, a sudden storm) as many times as they want. Decart just raised $300 million with Toyota and Nvidia backing it, so big names see this as a cheaper road to safer cars.
TOOLS ON OUR RADAR
🖋️Outline Free and Open Source: A fast and collaborative knowledge base built for remote teams to document their internal processes and share company guides in a beautifully designed interface.
⚙️ *Make Freemium: A visual automation platform that allows you to connect thousands of different applications and APIs to create complex workflows without writing a single line of code.
📧 Google CC Free: An experimental AI productivity agent from Google Labs that lives directly inside your email inbox to act as your personal executive assistant. It connects to your calendar and cloud files to automatically generate a personalized morning briefing organize your daily schedule and draft email replies.
🎬 PlayPhrase Freemium: A unique search engine that lets you type in any quote or phrase to instantly find matching video clips from thousands of movies and television shows.
TRENDING
MIT study: leaning on AI to spot fake news can backfire In a test of 67 people over four weeks, folks were 21% better at catching fake news with an AI helper. But once the AI was gone, their own skill dropped 15 points below where they started. Researchers call it the "AI dependency paradox."
Nvidia chips now help run Apple's Private Cloud Compute Apple is moving some Apple Intelligence tasks onto Google Cloud servers running Nvidia chips, using "confidential computing" that keeps your data locked even while it's being worked on, so the cloud operator can't peek.
Claude's Managed Agents can now run on a schedule Anthropic added two things: AI agents that run on their own on a set schedule (like a nightly data sync) and a safer vault for passwords and keys, so agents can log in to other tools without you handing over the secret.
Google's Gemini 3.5 Live Translate handles 70+ languages in near real time The new audio model listens as you talk and speaks the translation back just a few seconds later, keeping your tone and pace, with no setup. It's live now in the Google Translate app on Android and iOS.
Warner Music buys AI attribution startup Sureel Warner picked up Sureel, whose "AI DNA" tech tracks when AI models use an artist's music, voice, or likeness, so the artist can get credit and a cut. Sureel will keep running on its own.
Crawford CTO warns AI could weaken insurance talent pipelines A tech leader at claims giant Crawford warns that letting AI handle simple claims strips out the entry-level work that taught new adjusters the job, which could leave the field short on skilled people down the road.
Satya Nadella says companies should manage AI agents like employees Microsoft's CEO says as agents take on real tasks, businesses should give them IDs, limit what they can touch, and audit their work like staff. He says he runs about 100 coding agents at once, and the mental load of steering them is "so high."
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
⚜ List the accomplishments you've forgotten or downplayed. See what they reveal about who you actually are.
Build a single-file HTML app with vanilla HTML/CSS/JS. The Forgotten Win Audit — examine downplayed accomplishments to reveal real identity. Persist to localStorage key 'forgotten_win_audit_v1'.
Aesthetic: warm parchment cream (#f0e6d2) with subtle paper noise overlay, museum/archive curator feel. Cormorant Garamond serif for headings and italic body, Crimson Text for body, Special Elite typewriter mono for stamps and labels, IBM Plex Mono for field labels. Deep espresso (#2a1f17) text, antique gold (#b8862f) accent, forest green (#3a5d3a) for forward-looking section, rust red (#a3271c) for stamps. Rotated "EXAMINED & ENTERED INTO RECORD" stamp, "EXHIBIT A" corner tag on form, "SPECIMEN №01" numbered win cards with verified-checkmark watermark.
Form: wins textarea (one per line, strong help text encouraging small/old wins), era dropdown (last year / 3-5 years / decade / lifetime / specific), self-perception dropdown (not doing enough / lost my edge / proud but undersold / searching / rebuilding / don't know who I am), optional "why kept quiet" textarea.
System instructions to the model: act as careful respectful curator examining wins like exhibit specimens. NOT a cheerleader. Clear-eyed witness. Surface real significance, name through-line, give honest evidence about who they actually are. Specific not generic. Return raw JSON: verdict (1-2 sentence stark italic-worthy headline truth), pattern (2-3 sentences naming through-line referencing actual wins), wins_analysis array (one per win: win + hidden_significance + skill_revealed + why_downplayed), identity_findings array (4 "who you actually are" statements with **bold-italic** on key phrases), evidence_against_imposter (2-3 sentences using specific wins to counter their self-perception, direct address), what_built_for (2-3 sentences on what this suggests they're built for), title_to_claim (one specific earned identity title drawn from their pattern).
Render: dark verdict card with rotated stamp and large italic headline truth in quotation marks. Gold-left-bordered pattern card. Section dividers with ⚜ symbols and decorative side lines. Win cards as numbered specimen cards with italic win text + 3-row analysis grid (Significance/Skill revealed/Why hidden) with verified-checkmark watermark. Bordered "Who You Actually Are" card with ⚜ symbol per finding and bold-italic key phrases highlighted in rust red. Dark-mode evidence-against-imposter card with gold accent. Green-bordered "what you're built for" card. Centered final "title to claim" card with double-border glow effect and large italic quote.What this does: Type your forgotten wins one per line — the small ones, the old ones, the things nobody applauded. Pick the era under examination, your current self-perception, and why you've kept them quiet. Get back a museum-style examination: a stark verdict on what the audit reveals, the through-line pattern across your wins, each win analyzed as a "specimen" (significance / skill revealed / why hidden), four "who you actually are" statements drawn from the evidence, counter-evidence against your imposter using your specific wins, what you appear to be built for next, and one earned title to claim. Saves to localStorage.
What this looks like:
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Translate live speech across 70+ languages just a few seconds behind the speaker, keeping their tone, pitch, and pace (Gemini 3.5 Live Translate).
❌ Still Can't: Be trusted on its strongest public model for hacking, biology, or chemistry questions. Anthropic blocks those on Fable 5 and routes them to Opus 4.8.
✅ AI Can Now: Work on its own through a long task, finishing a coding job in a day that would've taken a team two months (Claude Fable 5, per Stripe).
❌ Still Can't: Build flawless interactive video worlds. Decart's Oasis 3 makes hours of lifelike driving but still ships "with some caveats."
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
A German rom-com where a buttoned-up scientist agrees to spend three weeks living with a humanoid AI companion named Tom (Dan Stevens, speaking German), programmed by an algorithm to be her ideal partner, in exchange for funding for her research. Director Maria Schrader treats every conversation between them like a Jane Austen novel rewritten for the age of personalized AI. Won the German equivalent of Best Picture and was Germany's Oscar submission, and almost nobody outside Europe saw 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.
What Replaces Roundup?
The next agricultural transition may not be bigger tractors. It may be autonomous robots replacing herbicides entirely. Greenfield Robotics is building commercial systems designed for that future.
Greenfield Robotics is Testing The Waters under tier 2 of Regulation A. No money or other consideration is being solicited, and if sent in response will not be accepted. No offer to buy the securities can be accepted and no part of the purchase price can be received until the offering statement filed by the company with the SEC has been qualified by the SEC. Any such offer may be withdrawn or revoked, without obligation or commitment of any kind, at any time before notice of acceptance given after the date of qualification. An indication of interest involves no obligation or commitment of any kind. “Reserving” shares is simply an indication of interest. There is no binding commitment for investors that reserve shares in this manner to ultimately invest and purchase the shares reserved of the company, or to purchase any shares of the company whatsoever.