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 at Work: Microsoft Caught Saying Addiction Is the Plan
TLDR: Microsoft got caught in an internal document saying it wants to make people "addicted" to its new AI assistant, Scout.
The Story: A leaked Microsoft strategy document, first reported by 404 Media and covered by Futurism, lays out a three-step plan for the company's new AI agent called Scout. Step one, in plain English: "Make people addicted." Scout is an always-on AI helper that lives inside Microsoft 365 apps like Teams, Outlook, and Word, and Microsoft announced it at the Build 2026 developer event. The document says more than 1,000 Microsoft staff, including CEO Satya Nadella, already use the tool every day. After the story broke, Nadella reportedly told employees he was "not sure" who wrote the document, even though the original report names the two Microsoft executives listed as its authors. Microsoft later sent a statement saying its real goal is "more time back, not more screen time."
Its Significance: This is one of the first times a big tech company has written down, in plain English, that addiction is the plan. We've heard the same thing before, just from the people who already quit. In 2017, former Facebook president Sean Parker told Axios that social media was built to give users a "little dopamine hit" so they'd keep coming back. The Netflix documentary "The Social Dilemma" featured dozens of ex-Google and ex-Facebook designers saying the same thing. Author Nir Eyal even wrote a whole book in 2014 called "Hooked" that teaches companies how to build habit-forming products, and Jonathan Haidt's 2024 book "The Anxious Generation" mapped out what those habits did to teen mental health. The playbook that hooked kids on TikTok is now aimed at your office, and Microsoft just said the quiet part out loud. Technology should be a way to save time, not steal it.
QUICK TAKES
The story: On June 4, OpenAI rolled out a new ChatGPT memory system called Dreaming V3, starting with US Plus and Pro users. The system runs in the background, learns from your past chats, and updates what it knows about you over time, so it understands your Singapore trip is over and you're back home.
Your takeaway: Memory used to be a list you had to manage. Now it's a profile that updates without you doing anything. Free users get it in the next few weeks after OpenAI cut the computing cost by about 5x. Open the new memory summary page once a month, look at what ChatGPT thinks it knows about you, and delete anything that feels off.
The story: Demis Hassabis, the Nobel Prize-winning head of Google DeepMind, told a Stanford audience that artificial general intelligence (AGI), meaning AI as smart as a human across most jobs, will show up around 2030, give or take a year. He said we're "in the foothills of the singularity" and society doesn't have long to get ready.
Your takeaway: Hassabis used to say 2030 to 2035. Last week he said 2029 to 2030. Now he says four years. The shrinking timeline matters more than the date itself. At the same time, AI critics point out today's top models score below 1% on the ARC-AGI-3 reasoning test that humans ace. Treat 2030 as one expert's bet, not a fact.
The story: Scientists at the University of Cambridge used machine learning to design a "super-antigen" that trains the immune system to fight whole families of viruses, not just one strain. A coronavirus version of the vaccine has cleared an early safety trial on 39 people, and a bigger 200-person trial is next.
Your takeaway: Today's vaccines react after a virus has already spread. This one looks at what every coronavirus needs to survive and trains your body for the whole family at once, even mutations that don't exist yet. If it holds up in the bigger trial, lockdowns and millions of deaths could be off the table for the next big outbreak.
TOOLS ON OUR RADAR
🖋️ FocusWriter Free and Open Source: A magnificently immersive writing environment that completely hides all your computer distractions leaving only a beautiful digital canvas so you can concentrate entirely on your words.
🏢 Marblism* Freemium: An innovative digital platform providing a complete team of artificial intelligence employees to seamlessly handle your business operations by automatically managing your email inbox scheduling social media posts and writing optimized articles completely in the background.
🖼️ Piwigo Free and Open Source: A delightfully elegant photo gallery platform that gives you absolute control to beautifully host organize and securely share your massive personal image collections with your family.
🍿 SMPlayer Free and Open Source: A wonderfully simple media playback software that remembers exactly where you stopped watching a movie and beautifully plays virtually every single video format in existence.
TRENDING
Nvidia Drops Nemotron 3 Ultra, an Open AI Model Built for Agents That Run for Hours — Nvidia released a 550-billion-parameter open model with a hybrid Mamba-Transformer design, built for AI agents that plan, use tools, and reason across many steps. Nvidia says it runs up to 6 times faster than other open models its size and handles up to 1 million tokens of context.
Amazon's AI-Generated Animated Series "Punky Duck" Killed After Two Days of Backlash — Emmy-winning director Jorge Gutierrez agreed to make an AI cartoon for Amazon's new GenAI Creators' Fund. After two days of online outrage, he dropped out, posting that "actions speak louder than words." The cancellation is the latest sign that audiences are pushing back hard on AI in entertainment.
Anthropic Says Claude Now Writes 80% of Its Own Code, and Humans Are Slowing Things Down — In a new report called "When AI Builds Itself," Anthropic said Claude writes more than 80% of the code in its own codebase, and its engineers ship 8 times more code per quarter than they did in 2024. The company warned this trend points toward "recursive self-improvement," where AI designs its own successor.
Anthropic Calls for a Global Pause on AI, Days Before Its Own IPO Filing Gets Talked About — Anthropic said the world should hit pause on the most powerful AI systems before they slip out of human control. They also confidentially filed for an IPO this year, just raised $65 billion at close to a trillion-dollar valuation, and want a slowdown only if every rival stops at the same time. The company asking everyone to slow down is also sprinting to cash in.
Brookhaven Lab Intern Uses AI to Train Real Humanoid Robots — Jasmin Lin, a master's student in bioinformatics, spent her internship at Brookhaven National Laboratory connecting "vision-language-action" AI models to physical robots that can see, follow instructions, and act. The work supports the Department of Energy's Genesis Mission to speed up scientific discovery with AI.
SoftBank's Masayoshi Son Says OpenAI's Next Model Is Being Designed by AI, Calls It a Sign of "Superintelligence" — Masayoshi Son, who runs SoftBank and is one of OpenAI's biggest investors, told CNBC that smarter-than-human AI is two years away, not the 10 he predicted in 2024. He also said he uses ChatGPT two to three hours a day because it's smarter than him.
Condoleezza Rice Says the US-China AI Race Will Decide the World Order — At the BNY Insite 2026 finance conference, former US Secretary of State Condoleezza Rice told business leaders the world is going through its biggest geopolitical shift in 80 years. She called AI the most important technology in that shift and said the US has to stay ahead of China for democracy to keep shaping the global rules.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
⛱️ Describe your situation. Get an Out Of Office message so good people leave you alone AND respect you more.
Build a single-file HTML app with vanilla HTML/CSS/JS. The Out Of Office Genius — OOO messages that protect time and earn respect. Persist to localStorage key 'ooo_genius_v1'.
Aesthetic: warm sand cream (#f0ebe2), soft golden vacation gradients, Playfair Display serif for headings and body italic, Inter for UI, DM Mono for labels. Burnt sienna (#a85020) accent. Rotated "⛱ AUTO-RESPONDER ⛱" stamp in header. White message cards with subtle ruled lines. Four-tab tone selector. Dark "set yourself up to actually be off" rules card with golden accents.
Form: reason input, out-from + back-on inputs, backup input, email-checking policy dropdown (not checking / once a day / mornings / emergencies / batch reply), audience dropdown (mixed / mostly internal / mostly clients / execs / creative peers), optional special context input.
System instructions to the model: write OOOs that (1) get recipients to leave them alone, (2) make recipients respect them MORE for protecting time. Best OOOs are confident, specific, warm, never apologetic. No "Sorry for any inconvenience" or "ASAP." Don't overshare medical/personal details. Match tone to audience. Return raw JSON: versions array (exactly 4: Professional / Warm / Cheeky / No Apologies — each with subject + body 3-5 sentences), strategy (2-3 sentences on why these work), short_version (under 100-char Slack status), setup_rules array (4-5 specific setup moves like "turn off mobile push"), predictions array (4 items with emoji + 1-line italic predicted reaction).
Render: four-tab selector for tones (Professional / Warm / Cheeky / No Apologies). Active version shown in printed-letter style card with subject line, body in serif, and meta row (Out / Back / Checking). Golden strategy callout. Slack-tagged short-version card. Dark numbered setup-rules card. White predictions card with emoji + italic line per reaction.What this does: Tell it why you're out, your dates, who covers urgent stuff, your email-checking policy, and your audience. Get back four versions in different tones (Professional / Warm / Cheeky / No Apologies) — each with subject line and complete body — plus the strategy explaining why they work, a short Slack/Teams status version, four setup rules to actually be off (turn off mobile push, etc.), and four predicted recipient reactions with emojis. Saves to localStorage.
What this looks like:
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Remember what you told it weeks ago and quietly update those memories as time passes. (ChatGPT Dreaming V3, OpenAI)
❌ Still Can't: Keep that memory accurate without a human checking what's stored. OpenAI added a review page because the system gets details wrong.
✅ AI Can Now: Write more than 80% of the code in its own developer's codebase. (Anthropic's "When AI Builds Itself" report)
❌ Still Can't: Pick its own research goals or design a fully new successor model on its own. Anthropic said that gap between AI today and AI building AI is still real.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
He, She and It by Marge Piercy - Book
A 1991 cyberpunk novel set in a near-future world dominated by corporate "multis," where a woman returns to a small free town and helps acclimate an illegal cyborg named Yod that was built to defend the town from corporate attack. Won the Arthur C. Clarke Award and is one of the best novels written about whether a person can be programmed and still be a person. Corporate AI governance about 30 years before anyone was talking about 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 happens when the S&P moves 3% during your commute?
We are living in volatile times. While you cannot control the state of international affairs, you can position your portfolio accordingly.
Liquid is one of the fastest growing trading platforms, allowing users to trade stocks, commodities, FX, and more 24/7/365 from their phone and computer.