In partnership with
7 Proven External Traffic Strategies
Most eCommerce brands running external traffic aren't scaling — they're just spending.
Wrong channels, no real attribution, and at the end of the month, still no clear answer to the question that matters: what actually drove revenue?
The brands getting it right aren't necessarily spending more. They've just stopped guessing. They know which channels pull weight on Amazon listings, which ones bleed budget, and why affiliate and creator traffic outperforms on ROI when it's set up correctly.
Levanta put together a free playbook breaking down 7 proven external traffic strategies — where each one works, where it falls apart, and what it takes to scale without it becoming a second job.
Inside you'll see how top brands are driving millions in off-Amazon revenue and why most channels underdeliver when brands don't know what to look for.
If you're serious about growing outside of PPC, this is worth 5 minutes.
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
Google's Top AI Scientist Says a Watchdog Should Screen AI, Even Google's Own
TLDR: The head of Google's AI lab wants the U.S. to build a new safety group that would test the most powerful AI before it comes out, and could even tell companies to slow down if things get risky.
The Story:
Demis Hassabis runs Google DeepMind, the lab behind the Gemini AI. This week he wrote a plan asking the U.S. government to set up a watchdog for AI. The group would check the strongest AI models before they reach the public. If a model looked dangerous, the watchdog could ask AI companies to pump the brakes across the whole industry.
The plan is specific. Companies would hand over their newest models up to 30 days before release, so experts could test them for risks like cyberattacks or help building bio and nuclear weapons. At first this would be by choice. Later, passing the test could become a rule before a model could launch in the U.S. Hassabis wants the group modeled on FINRA, the industry-funded body that watches Wall Street. He points to last month's sudden U.S. freeze on Anthropic's most powerful models as a sign the country needs clear rules, not rushed ones.
Its Significance:
The people building AI are now asking the government to police them, which almost never happens in tech. Hassabis thinks AI as smart as a person could be only a few years away, so he says the time to set rules is now. Tough, costly testing could be easier for a giant like Google to handle than for a small startup, so some worry it could box out smaller players in favor of the largest companies. This is an early look at who might get to decide which AI tools are safe enough to reach your phone, your job, and your kids.
QUICK TAKES
The story: Apple's watchOS 27 beta brings a smarter "Siri AI" to the Apple Watch, turning the old timer-and-weather helper into a chatbot-style assistant you can ask real questions and have back-and-forth chats with. Early testers call it promising but buggy, and the full version lands later this year.
Your takeaway: The watch on your wrist is about to go from step counter to something you can actually talk to.
The story: On the bank's earnings call Tuesday, CEO Jamie Dimon said AI has already shrunk certain roles at JPMorgan by 30% to 40%, though he added most of those workers were offered other jobs. The bank still had about 320,000 employees at the end of June, roughly flat, and Dimon says AI now runs almost a thousand tasks there, from fraud checks to note-taking.
Your takeaway: This is one of the clearest signs yet that AI is changing real jobs at a huge company, even when the total headcount hasn't dropped.
The story: Hugging Face CEO Clem Delangue says the real fight in AI is moving from a few giant top-tier models toward cheaper open ones that anyone can download and run for free. He notes companies often start on pricey AI services, then switch to open models as they grow, and open models are now used by about half the Fortune 500.
Your takeaway: The most expensive AI isn't always the one that wins, and cheaper open tools are quietly getting good enough for real work.
TOOLS ON OUR RADAR
🎙️ Granola Freemium: An AI notepad that turns your meetings into polished notes automatically, no recording bot needed since it just listens through your computer's audio.
📼 Longscribe Freemium: Paste a link from YouTube, TikTok, or a podcast and get a full text transcript with subtitles, no software to install.
📅 Calendly Freemium: Share one link and let people book time on your calendar directly, no more back and forth emails trying to find a time that works.
🔐 Proton Pass Free and Open Source: A privacy first password manager from the Proton team with built in email aliases, so you can protect your accounts without paying a cent.
TRENDING
OpenAI Says You're Probably Prompting ChatGPT Wrong - OpenAI put out a new guide for its GPT-5.6 model with a surprising message: stop writing long, detailed prompts and just describe what you want. In its own tests, leaner prompts gave 10% to 15% better results while cutting costs by up to two-thirds.
Perplexity's AI Assistant Now Remembers Your Past Work - Perplexity updated its Computer agent with a memory system called Brain that learns from your past tasks overnight, plus a way to build and publish a website in one click. The company says the memory feature made answers 25% more correct on tasks it had seen before.
Claude's "Personality" Changes With the Language You Use - Anthropic studied more than 300,000 chats and found its Claude AI shows different values depending on the model and the language, warmer in Hindi and Arabic, and blunter and more careful in English and Russian. The company says it doesn't yet know how much of that difference is a good thing.
Meta Wants to Rent Out Its Massive AI Computing Power - Meta is moving to sell access to its huge AI data centers, which would put it up against Amazon, Microsoft, and Google in cloud computing. This week it said it would spend $50 billion to expand one Louisiana site alone.
Spotify Adds a ChatGPT-Style Assistant for Music - Spotify Premium users can now chat with the app by voice or text to pick what to play and ask about their listening history. It's in beta for adults in the U.S., Ireland, and Sweden.
An Open-Source AI Agent Maker May Be Worth $1.5 Billion - Nous Research, the team behind the free, open Hermes agent, is in talks to raise at least $75 million at a $1.5 billion value, less than three months after its last round. Hermes has picked up around 214,000 stars on GitHub.
MIT Built AI That Makes Virtual Playgrounds to Train Robots - MIT researchers made a system called SceneSmith that uses AI to build detailed virtual rooms, like a garage or a kitchen, where robots can practice chores before trying them for real. The rooms hold up to six times more objects than older methods.
Playing Against AI Opponents Made Gamers More Social - University of Florida researchers found that after the game PUBG: Battlegrounds quietly added AI opponents, new players won more, stuck around about 50% longer, and teamed up with friends 28% more often. The lead researcher says the same idea could work at a job, with AI helping newcomers build confidence instead of replacing them.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
🪫Answer 8 honest questions. Get your burnout score, dimension breakdown, and a personalized recovery prescription.
Build a single-file HTML app using vanilla HTML, CSS, JS, and one Claude API call. Create Burnout Barometer — an 8-question burnout assessment with scoring, dimension breakdown, and recovery prescription. Save all results to localStorage key 'burnout_barometer_v1'.
Aesthetic: off-white paper background (#f5f2ec), stark black typography, red (#c0392b) as the danger accent. Typography: Syne 800 for headings with tight letter-spacing, Lora italic for prescriptions, Syne Mono for labels. Clinical but human.
Eight questions covering: Exhaustion, Detachment, Efficacy, Recovery (inverted), Autonomy (inverted), Resentment, Meaning (inverted), Outlook (inverted). Each has a 1–5 scale with descriptive endpoints. Selected buttons color-code: red for 4–5, amber for 3, green for 1–2. A progress bar fills as questions are answered. Submit button enables only when all 8 are answered.
Calculate a 0–100 burnout score. Invert the appropriate dimensions before scoring. Send score and all 8 dimension values to Claude API requesting raw JSON: label, description, optional warning (if score ≥ 70), immediate actions array, structural changes array, reframe string.
Render: animated count-up score display with color (red/amber/green), 8 dimension bars with color-coded fills, prescription sections for this week and structural changes, a reframe quote at the bottom. Store each assessment with date, score, label, and parsed data. Show past assessments as a history list with color-coded scores. Make it work in a single HTML file.What this does: Eight questions covering exhaustion, detachment, efficacy, recovery, autonomy, resentment, meaning, and outlook. Each answered on a 1–5 scale. Claude scores you 0–100, names your burnout stage, breaks down all eight dimensions as progress bars, flags an urgent note if your score crosses 70, and delivers two sets of prescriptions: what to do this week and what structural changes to make longer-term. It also ends with a single honest reframe — something true and helpful about being at that stage. Every assessment saves to localStorage with the date and score, so you can track whether things are actually getting better.
What this looks like:
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Build detailed virtual rooms, complete with real-world physics like weight and friction, so robots can practice chores in a digital space (MIT's SceneSmith).
❌ Still Can't: Reliably pull those chores off in the real world. When MIT tested robot plans across 100 of these virtual rooms, the robots often failed at simple tasks.
✅ AI Can Now: Hold a real back-and-forth conversation on your wrist through the Apple Watch's new Siri AI.
❌ Still Can't: Do it dependably yet. Early testers say the watch assistant often stalls, times out, or cuts you off mid-question.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
Elysium by Jennifer Marie Brissett - Book
A short, structurally daring 2014 debut novel about a corrupted AI computer program etched into Earth's atmosphere that is trying to tell the story of two people, Adrianne and Antoinette, whose relationship and sex and eras keep changing every time the program glitches and starts over from a new angle. Won a Philip K. Dick Special Citation and was a Locus Best First Novel finalist. A broken AI trying to preserve human memory.
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
Moda is an AI design product where you prompt what you need, get a complete on-brand design, and edit every element on a full canvas.
Our viral launch hit 4.4M views in days, tens of thousands signed up, and executives at major finance and tech companies now use it.