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
OpenAI Spent $6.5 Billion to Build a Speaker That Acts Alive
TLDR: OpenAI's first gadget will be a screen-free smart speaker that talks like a person, moves parts of itself to seem alive, and watches the room with a camera.
The Story:
OpenAI's first piece of hardware won't be a phone. It's a smart speaker with no screen. You can pick it up and carry it from room to room because it runs on a battery. Bring it to the kitchen for cooking help, then move it to the bedroom to play music. It talks using GPT-Live, OpenAI's newest voice model, which can listen and speak at the same time.
The speaker has a camera and other sensors, so it can take in what's around it. It also has small moving parts, built to make the machine feel alive instead of like a plain box. OpenAI paid $6.5 billion last year to buy io, the hardware company started by former Apple designer Jony Ive. The plan is to show it off this year and sell it in 2027. One catch: Apple just sued OpenAI, claiming it stole trade secrets, so the date could change.
Its Significance:
A speaker that carries a camera and listens all day raises a big privacy question. It's built to learn about you over time, even reading your emails to get to know you better. That's handy when it hands you the right reminder. It's also a lot of personal information sitting in one always-on device in your home. Alexa has already been embroiled in a few privacy-related scandals with Amazon’s speakers.
It also shows where AI is inevitably heading. It's moving off your screen and into the room with you. If it works, asking for help could feel less like typing and more like chatting with someone in the kitchen. Whether people want a machine that acts alive with a camera on them remains to be seen. Meta is betting they will, only in glasses form.
QUICK TAKES
The story: Justin McLeod, who founded Hinge, raised $18 million for a new company called Overtone, and even Hinge's owner Match Group chipped in. Overtone drops swiping and profiles; instead an AI listens to you talk and then makes a few hand-picked introductions.
Your takeaway: The man who taught everyone to swipe now says swiping doesn't work. If this catches on, AI moves from matching you to actually deciding who you meet.
The story: Star Wars creator George Lucas said AI makes movies "much easier" to make and compared fighting it to clinging to a horse and buggy after cars showed up. He also said AI could help flag fake videos and tag where they came from, something he thinks people can't do well on their own.
Your takeaway: A director famous for pushing movie tech is backing AI while much of Hollywood pushes back. When the guy who built Industrial Light & Magic calls AI the future, studios listen.
The story: Google DeepMind CEO Demis Hassabis wrote that human-level AI is "probably only a few short years away" and could matter more than fire or electricity. He also warned the risks are growing fast and asked the U.S. to build a new group that tests the most capable AI models before they're released.
Your takeaway: When one of the top AI scientists says this beats fire, it's worth a pause. His fix, a testing body before release, hints at how AI might get regulated soon. Many claims from tech CEOs have had to be walked back, though Demis being one of the most respected scientists in the field carries more weight.
TOOLS ON OUR RADAR
📊 ReviewTactic Freemium: Connects to your Google Business Profile in one click and drafts replies to customer reviews automatically, useful for small business owners but not a casual pick for most readers.
🔍 Adviserry Paid: Builds a searchable AI knowledge base from your content, aimed at teams organizing internal documentation rather than a beginner's everyday tool.
🖥️ askbowtie Freemium: Uses AI agents to monitor websites for changes, a developer and ops tool that assumes technical comfort with automation.
🗣️ Talkory.ai Paid: Compares how different AI models respond to the same prompt, useful mainly for people evaluating LLMs rather than everyday users.
TRENDING
Daydreaming Algorithm Helps AI Remember What Matters — Researchers built an algorithm that copies how sleep works, letting an AI memory network learn new things and clear out junk memories at the same time. A new version handles messy real-world data, like very dark or very bright images, without wiping the good memories. The goal is AI memory that's more reliable and uses less energy.
Can AI Build a Jet Engine? JARVIS Challenge Tests AI Copilots in Tough-Tech Engineering — MIT gave undergrad teams four weeks to design, build, and safely fire a small jet engine using AI as their main engineering partner. One team pulled it off, and professors said AI sped up the work but human judgment and hands-on skill still decided the winner. Their takeaway: in the AI era, real training matters more, not less.
Google Faces Another AI Training Lawsuit From Major Publishers — Big publishers like Hachette, Cengage, and Elsevier, plus author Scott Turow, sued Google, saying it trained its Gemini AI on their books without permission. The suit claims Google used books it was only allowed to show as short snippets, then stripped copyright info to hide it. A New York judge will weigh in, after earlier California rulings leaned toward AI companies.
At Last, a Good Reason to Buy an AI PC: Reining in Runaway Token Bills — Analyst firm Gartner says businesses can cut sky-high cloud AI bills by running smaller AI models right on desktop AI PCs. Those machines have chips fast enough to handle lighter jobs, keeping work off pricey cloud servers. Gartner suggests companies start testing this once newer AI PCs arrive in 2027.
South Korea to Launch Universal Basic AI Chatbot — South Korea's government is paying private companies to build a free AI chatbot every resident can use, plus an AI agent for handling government services. The country will hand winning bidders up to 256 Nvidia B200 GPUs, and the deal runs through 2031. The aim is a home-grown service so people can't be cut off by decisions made in other countries.
New York Becomes First State to Halt Datacenter Buildouts — New York Governor Kathy Hochul paused permits for large data centers, 50 megawatts or bigger, for up to a year while the state writes new rules. The worry is that AI data centers push up electric bills, drain water supplies, and hurt air quality. It's the first statewide freeze of its kind and could become a model for other states.
OpenAI Hides Codex Agent Instructions Behind Encryption, Leaving Developers in the Dark — OpenAI now scrambles the messages its Codex coding agents pass to each other, so developers can't read what one agent tells another. Coders worry this makes it harder to debug and check what the AI actually did. OpenAI hasn't said why, but some think it's to stop rivals from copying how its system works.
TRY THIS PROMPT (copy and paste into Claude, ChatGPT, or Gemini)
🔎 Walk through your plot like you're briefing a detective. Find the logical gaps and continuity risks before your readers do.
Build a single-file HTML app with vanilla HTML/CSS/JS. The Plot Hole Hunter — describe your story's plot, get logical gaps and continuity risks flagged, detective-case-file style. Persist to localStorage key 'plot_hole_hunter_v1'.
Aesthetic: noir dark (#100e0b), amber (#d4a24c) primary with amber glow top-left, faint red glow bottom-right. Special Elite (typewriter font) for headings and form-prompt/buttons, Crimson Pro serif for body/issues, JetBrains Mono for labels. Dashed borders throughout for a "case file" feel. Severity colors: critical=red (#c0392b), moderate=amber (#d4a24c), minor=steel blue (#6b8ba4).
Form: large plot-summary textarea (encourage a synopsis/outline/beat sheet), format dropdown (novel / short story / screenplay / TV series arc), optional specific-worry text input.
System instructions: meticulous plot-logic detective hunting genuine holes — continuity errors, unearned motivation shifts, timeline inconsistencies, world-rule violations, over-convenient coincidences, information problems (knowing/not-knowing things they shouldn't). Work ONLY from what's actually described, don't invent unstated details, but flag where the described plot doesn't hold together. Rate severity critical/moderate/minor. Address any stated worry directly. No em dashes. Return raw JSON: case_summary (2-3 sentences, detective overview), holes (3-6: title, severity, issue, why_it_matters, fix with **bold** core fix), loose_threads (2-4 dangling setups, empty array if none), verdict (2-3 sentences, **bold** the priority fix). Be a fair investigator, not a nitpicker inventing problems.
Render: gradient "case summary" hero card. "Evidence: holes found" section — cards left-bordered by severity color, showing title + severity badge + issue + a mono "reader impact" line + a dashed-divider italic fix (bolded core). Amber "still-open threads" card with em-dash list items. Red-bordered "VERDICT" card with bolded priority. Copy findings + archive of closed cases keyed by plot snippet + format.What this does: Summarize your plot, pick the format, and optionally name a specific worry. It investigates only what you actually described, no invented details, and flags genuine holes: continuity errors, unearned motivation shifts, timeline problems, world-rule violations, and coincidences doing too much work. Each finding is rated critical, moderate, or minor, with the specific issue, why it costs reader trust, and a concrete fix. You also get a list of loose threads left dangling, and a closing verdict on how serious it all is together with the one fix worth tackling first. Saves each case to localStorage.
What this looks like:
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Help students design and safely fire a working small jet engine in just four weeks, with AI copilots as the main engineering partner.
❌ Still Can't: Beat human hands at the physical build. At MIT, making the parts, not the AI-assisted design, was the slowest step.
✅ AI Can Now: Learn new patterns while scrubbing out false "memories" at the same time, even on lopsided real-world images.
❌ Still Can't: Do that scrubbing safely on its own. Let the cleaning run too long and it erases correct memories too.
RECOMMENDED LISTENING/READING/WATCHING
Cascade Failure by L.M. Sagas - Book
The 2024 debut novel of the Ambit's Run series, about the ragtag crew of an old Guild ship captained by a curious AI called Eoan, whose relentless drive to keep learning has taken them further and further out into the Spiral than any other captain will risk going. When they stumble across a distress call from a dead planet, they end up dragged into a much bigger conspiracy about who exactly the interstellar Trust is willing to sacrifice. Pitched as Firefly meets Murderbot and one of the most interesting AI captain characters since Ancillary Justice.
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.