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

Estonia Becomes the First Country to Give AI Agents Their Own ID

TLDR: Estonia plans to become the first country to give AI agents their own official digital ID codes, so an AI helper can act for you without holding all your passwords and data.

The Story:
Right now, when you let an AI assistant do a task for you, it usually borrows your whole login. It gets your accounts, your files, and your money access, all at once. Estonia wants to change that. Prime Minister Kristen Michal backed a plan on June 17 to create what officials call "AI ID codes," separate digital identities just for AI agents. With one, an AI could be told it may only view your data, only prepare a document, or only approve payments up to a set limit. Estonia already runs a state-backed e-ID system that lets people do almost all government tasks online, so it's building this on top of something that works. The country says it wants every AI action to be checkable and tied to a clear owner who is responsible.

Its Significance:
AI agents are about to handle real tasks: filing forms, moving money, booking things. If something goes wrong and the AI had your full login, it's hard to say who did what. Giving the AI its own ID means you can hand it a small, fenced-in job instead of the keys to everything.The downside is that digital ID is a favored surveillance tool for governments around the world and would create a permanent, trackable record of everything we do online.

Newsletter Launch Cohort
Newsletter Launch Cohort
Launch your professional newsletter in six weeks alongside a small group. Six live Saturday classes(virtual) starting July 18, plus a shared AI content engine and a community. You leave with a comp...
$250.00 usd

QUICK TAKES

The story: Researchers from the Broad Institute and Wyss Institute trained an AI on 38,650 lab-tested molecules, then used it to scan about 6 million compounds for ones that could kill a hard-to-treat STI. The AI found two that work in brand new ways, and in mice and lab models both wiped out the bacteria without triggering resistance.

Your takeaway: Gonorrhea has slowly beaten almost every drug we throw at it, with more than 600,000 US cases a year. An AI just did in a short stretch what would take humans years of lab work, pointing to a faster path for finding drugs when bugs stop responding to the old ones.

The story: Researchers trained an AI to look at standard ECG heart scans and estimate the heart's biological age, which can differ from your real age. When the heart "tested" seven years older than the person's actual age, the risk of death and major heart problems jumped sharply.

Your takeaway: Your birth certificate says one number, but your heart might be living at a different one. A quick scan you already get at the doctor could soon flag trouble years earlier, while there's still time to act.

The story: A researcher tested ChatGPT with hiring-style questions and found it carries age bias, often nudging against workers over 45. The bias gets baked in through the data these tools learn from, which means AI can speed up old discrimination instead of fixing it.

Your takeaway: If a company uses AI to screen job applicants, an older candidate could get filtered out before a human ever sees them. The tools look neutral, but they aren't, and current laws don't fully cover this yet.

TOOLS ON OUR RADAR

🚀 Dashy Free and Open Source: A highly customizable homepage that allows you to organize all of your favorite websites and personal applications into one beautiful and easy to navigate digital dashboard.

🔍 SearXNG Free and Open Source: A privacy respecting search engine that aggregates results from many different sources while keeping your search history and personal data completely anonymous.

🔔 Ntfy Free and Open Source: A simple service that allows you to send push notifications directly to your phone or desktop from any application without needing a complex setup process.

🎵 Navidrome Free and Open Source: A fast and lightweight music server that lets you listen to your own personal audio collection from any device using any compatible music player.

TRENDING

AI Is Changing What Bosses Want From New MBA Grads — A new Cornell report says companies are automating the entry-level work MBAs used to do, so new hires will be expected to supervise AI like a junior employee and catch its mistakes. Skills that can't be automated, like storytelling and persuasion, are becoming the real edge.

A Robotic Kestrel Is Teaching Engineers How Birds Beat Rough Air — Researchers at RMIT and the University of Bristol built a robot bird that copies a kestrel's moves to learn how it stays steady while hovering in wind. The tricks they found could help small delivery and rescue drones fly in turbulence that grounds them today.

Your Smartwatch May Soon Notice When You're Struggling — University of Ottawa researchers built an AI system that reads heart rate, voice tone, and text to spot emotional distress early and offer support, before you even ask. In tests it scored higher than standard chatbots on empathy and personalization.

New "Skin" Lets Robots Feel Heat and Pressure at Once — A team at Seoul National University made a thin artificial skin that senses temperature and pressure in one device, much like human skin. Using it, a robot identified 20 everyday objects by touch with accuracy close to a human hand.

AI Spots a Hidden Heart Death Risk That Normal Tests Miss — UC Berkeley researchers trained AI on hundreds of thousands of ECGs and found a warning sign for sudden cardiac death, even in hearts that looked normal. The signal hadn't been described in medical books before, and it could help doctors flag at-risk patients earlier.

A Top Hedge Fund Is Building Its Own AI Lab — Millennium Management is setting up an in-house AI lab, going live in the coming weeks, to test new AI tools early and partner with AI firms. It's a sign that big finance now treats building AI know-how as a core part of staying competitive.

TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)

📈 A market dashboard for your watchlist: live prices, day moves, 5-day trends, the day's leaders and laggards, and where each name sits versus its 50-day average.

Build a single-file HTML dashboard (vanilla HTML/CSS/JS) called Stock Watch that shows a live market overview for a user-defined watchlist. Pull LIVE quotes via the Anthropic API's web search tool (there's no stock-price connector, so search is the data source).

At the top of the script, define an editable WATCHLIST array of tickers and the index symbols to track. On load, POST to https://api.anthropic.com/v1/messages with model 'claude-sonnet-4-20250514', max_tokens 1500, and tools: [{ "type":"web_search_20250305", "name":"web_search" }]. In the user message, instruct the model to look up the latest available price and day percent change for each watchlist ticker and for the major indices (S&P 500, Nasdaq, Dow, Russell 2000), and return ONLY raw JSON (no preamble/markdown): { as_of, market_status, indices:[{name,val,chg}], stocks:[{tk,name,price,chg,above50:boolean}] }. Assemble the model's text output from the content blocks (filter type==='text', join), strip ```json fences, JSON.parse. Wrap in try/catch with a clear error state, show a loading state, and note that quotes may be delayed and are not real-time exchange data. (5-day sparkline arrays can be approximated or requested too; if absent, hide the sparkline gracefully.)

Aesthetic: financial-terminal dashboard. Near-black (#080b10), cyanblue accent (#23c4b6 / #2e8de0) with top radial glow, JetBrains Mono for ALL numbers/tickers, Inter for text. Green (#2ecc71) up / red (#ff5a5a) down everywhere. Pulsing "Market open" status dot.

Render: topbar (brand mark + title + status pill), a 4-up index strip (name, value, colored / % change), a 4-stat row (symbols on watchlist, average move colored, top gainer ticker, green/red count). Then a two-column grid: left = Watchlist table (Symbol with company name, Last price, inline SVG 5-day sparkline colored by direction, Day % bold-colored, and an above/below 50-DMA chip); right = "Today's Movers" split into Leaders (top 3 gainers) and Laggards (top 3 losers) from the list, plus a "Watchlist Breadth" green/red split bar. All charts hand-built with inline SVG/divs, no libraries. Footer: "Market data for informational purposes · not financial advice." Compute stats, movers, and breadth client-side from the fetched data.

What this does: Tracks a watchlist of stocks at a glance. A top index strip shows the S&P 500, Nasdaq, Dow, and Russell 2000 with their day changes, then a stat row summarizes the list (symbols tracked, average move, top gainer, advancing vs declining count). The main watchlist table lists each symbol with its last price, a color-coded 5-day sparkline, the day's percent change, and an above/below 50-day moving-average chip for quick trend read. A side panel surfaces the day's leaders and laggards from your own list plus a green/red breadth bar. Pulls live quotes each time it loads via the API's web search tool. Market data only, not financial advice.

What this looks like:

WHERE WE STAND(based on today’s news)

AI Can Now: Read a routine heart scan and flag a risk of sudden cardiac death that today's standard tests miss entirely.

Still Can't: Judge a job candidate without absorbing the age bias buried in its training data.

AI Can Now: Sense temperature and pressure through a single artificial skin and tell apart 20 everyday objects by touch.

Still Can't: Be trusted with your full digital identity, which is exactly why Estonia is building separate ID codes to fence in what an AI agent is allowed to do.

Beginners in Stock Trading

Beginners in Stock Trading

Unique blend of beginner trading education and daily trading news to learn from the greatest stock traders in history.

FROM THE WEB

RECOMMENDED LISTENING/READING/WATCHING

We Are Satellites by Sarah Pinsker - Book

A 2021 near-future novel about a family of four in a country where everyone is suddenly getting a brain implant called the Pilot, which lets people multitask and supposedly gives users a leg up at work and school. Pinsker tells the whole story from inside the household, following each family member as they make different choices about whether to get the chip and how to live with each other afterward.

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.

Postgres Didn't Fail You. Your Architecture Did.

Adding a second database was supposed to fix things. Now you manage sync, drift, and pipelines on top of queries that are still slow.

TimescaleDB extends Postgres instead. Hypertables, 95% compression, continuous aggregates. One database. No pipeline.

Reply

Avatar

or to participate

Keep Reading