In partnership with
How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads
For its first CTV campaign, Jennifer Aniston’s DTC haircare brand LolaVie had a few non-negotiables. The campaign had to be simple. It had to demonstrate measurable impact. And it had to be full-funnel.
LolaVie used Roku Ads Manager to test and optimize creatives — reaching millions of potential customers at all stages of their purchase journeys. Roku Ads Manager helped the brand convey LolaVie’s playful voice while helping drive omnichannel sales across both ecommerce and retail touchpoints.
The campaign included an Action Ad overlay that let viewers shop directly from their TVs by clicking OK on their Roku remote. This guided them to the website to buy LolaVie products.
Discover how Roku Ads Manager helped LolaVie drive big sales and customer growth with self-serve TV ads.
The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.
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 Wrote the Intel Report. Troops Almost Boarded the Wrong Ship
TLDR: A US military analyst used a chatbot to write an intelligence report about a Chinese ship's cargo, the report was wrong, and planes were already in the air before anyone caught the mistake.
The Story:
An analyst at a US special operations command asked a chatbot to look at reporting about what a China-flagged ship was carrying. The bot blended public information with secret signals intelligence and decided the ship held material tied to nuclear weapons. The analyst then used AI again to turn that finding into a standard intelligence report, the kind commanders act on. And they did act. Military planes went up, and armed service members got ready to board the vessel. Officials took a closer look at the report right before the operation and found the AI had made the cargo up.
Its Significance:
AI is good at a lot of things right now, and the list keeps growing. It reads medical scans, writes usable code, designs proteins, and turns messy audio into clean text. Two stories further down this issue show that: an MIT system lines up X-rays with 3D scans during surgery in under a second, and a new speech model cut its error rate on short commands from 20.6% to 6.8%.
Here's where it’s not always great. When AI pulls from several sources at once and has to decide what's true, it will sometimes invent an answer and state it with total confidence. That's what happened with the ship. The report read like every other report, which is the problem. A person has to check the source of anything an AI tells you before someone acts on it. The higher the stakes, the more that matters, and it's hard to think of higher stakes than a boarding party climbing onto another country's ship. There's a possible future fix to this covered in yesterday’s lead story: Jev, a new model from ChatGPT co-inventor Diogo Almeida, doesn't write sentences at all. It picks from options you define ahead of time and hands back a confidence number, so in a case like the ship it could return something like "62% chance of nuclear material" instead of a persuasive paragraph stating it as fact.
QUICK TAKES
The story: During procedures like opening a blocked artery, doctors steer a tube through your body using flat X-rays, and those flat images make it hard to tell where the tool is. MIT's system, called xvr, trains on thousands of fake X-rays built from your own CT or MRI scan, then matches the live X-ray to that 3D picture with sub-millimeter accuracy in seconds.
Your takeaway: This is the other side of the lead story. Same technology, very different job. The MIT team trained one model per patient instead of one model for everybody, tested it on records from five hospitals covering kids and adults, and beat other AI methods by roughly ten times. Narrow job, checkable answer, measured result. That's where AI is winning right now.
The story: Anthropic confirmed to TechCrunch that it runs a wet lab in the Bay Area, meaning a lab with real biological material, not just simulations. Its head of life sciences, Eric Kauderer-Abrams, told Reuters the final test in biology is still real lab work, and the company says the focus is basic biology rather than drug discovery.
Your takeaway: AI companies have been saying for years that models will help cure disease. Running your own lab is what that looks like. Anthropic also bought a stealth biotech called Coefficient Bio in April and just opened a program giving vetted bio researchers access to its strongest models.
The story: Particle6 Group put its AI-generated "actress" Tilly Norwood in front of 75 journalists at once to promote a film called Misaligned. In a recorded sit-down with Piers Morgan and actor Tom Conti, Norwood failed to follow a simple question about whether her co-stars were real, then broke into Chinese for more than ten seconds before apologizing for a hiccup.
Your takeaway: This was a publicity stunt from the start. Putting an AI character on a talk show is like booking Buzz Lightyear for a late-night interview, and the coverage is the whole point. SAG-AFTRA is annoyed, reporters wrote it up, and mission accomplished for Particle6. There's nothing wrong with a company drumming up press in creative ways, as long as you see the product for what it is.
TOOLS ON OUR RADAR
📧 Thunderbird Free and Open Source: Manage multiple email accounts in one client with built-in encryption and spam filtering, and RSS feed, no self-hosting needed, just download and set up your accounts.
✍️ Scrivener Paid: Organize a long manuscript with a corkboard, outline, and full manuscript view in one place, one time purchase around $59.99, no subscription.
🤝 Freshsales Freemium: Manage contacts, deals, and pipelines for up to 3 users free, with built-in phone and email logging included at no cost.
🔐 Proton Pass Freemium: Store passwords and generate email aliases with end-to-end encryption, unlimited on the free plan across one device.
TRENDING
xAI doubled the accuracy of its transcription model and kept the price the same — Grok Voice Transcribe 2.0 still costs $0.10 per hour of recorded audio. xAI says the biggest gain came in other languages, where the error rate on short spoken commands across 19 languages dropped from 20.6% to 6.8%. Atlassian is the launch customer and uses it on Loom videos.
A new test asks whether AI can build a working robot — Researchers at Harvard and Georgia Tech released RLE-Bench, a free 48-task exam for coding AI. The catch for the agents is physics. One task asks for a rolling base that holds several robot arms, and code that looks fine on screen tips over the moment you add weight.
Claude Code will now run several coding jobs at once — Anthropic rebuilt Projects so one conversation splits into threads, each its own cloud session on its own copy of the code, still running after you shut your laptop. The Register points out the bill: every thread counts as a full session, so plan limits arrive faster. It's in beta for some Pro and Max subscribers. I would avoid using this feature for light tasks.
Google Home opened up to outside AI agents — Any assistant that speaks MCP, including ChatGPT and Claude, can now read your device history and flip your lights. Ask what the porch camera saw after school, or how many loads of laundry ran last week. It's US-only for now, limited to the $20-a-month Premium Advanced tier, and Google blocks sensitive moves like unlocking doors.
TRY THIS PROMPT (copy and paste into Claude, Grok, ChatGPT, or Gemini)
🏷️ Name what you're shopping for. Find out when it goes on sale, grounded in real current buying guides, not a guess.
Build a single-file HTML app with vanilla HTML/CSS/JS. The Best Time to Buy Calculator — name a product category, get real seasonal sale pattern research via live web search.
Aesthetic: dark near-black green (#0f140f), kelly green (#5cc26e) primary with a green glow top-left, gold (#e0b458) for the discount stat with a glow bottom-right. Inter for headings/body, JetBrains Mono for labels.
Form: single product-category text input (encourage a type of thing, not a specific model).
Technical: call the Anthropic Messages API WITH tools:[{type:'web_search_20250305', name:'web_search'}] (works automatically), system prompt as a savvy consumer shopping researcher who searches for CURRENT, real published buying guides and retail pattern reporting on when this product category typically goes on sale, grounding timing windows and reasoning in real current sources rather than generic assumptions, honest about discount ranges only when confidently found via search. Return raw JSON: product, typical_discount (general range if confidently found, empty string otherwise), best_windows (2-4: window + why, specific to the category), worst_time (1-2 sentences on peak-price timing to avoid), if_you_need_it_now (1-2 sentences of practical advice: price tracking, open-box, negotiating, older models).
Render: a gradient "typical discount when timed right" hero card with the product name and discount figure. A "best windows to buy" list of green-bordered cards (window + why). A red "avoid buying" card. A gold "need it now?" card with practical fallback advice.What this does: Enter a product category, a mattress, a laptop, patio furniture, and it searches the live web for real, current retail pattern reporting on when that category typically drops in price and why, model-year clearances, seasonal markdowns, big sale events, whatever applies to that specific thing. You get two to four real timing windows with the reasoning behind each, a typical discount range when one's confidently found, a clear warning on when prices tend to peak so you know what to avoid, and practical advice for if you can't wait for the ideal window, price tracking tools, open-box options, negotiating. It's grounded in current search results rather than stale assumptions about how retail sales work.
What this looks like:
Your traffic is fine. Your signups aren't.
Visitors land and leave, and "looks fine to me" isn't a diagnosis. SureThing audits SEO, speed, mobile, and messaging against the page, then ranks the fixes by impact.
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Match a live surgical X-ray to your own 3D scan with sub-millimeter accuracy in seconds
❌ Still Can't: Combine several intelligence sources and reliably tell you when it doesn't know the answer
✅ AI Can Now: Transcribe dozens of languages in one recording and follow a speaker who switches mid-sentence
❌ Still Can't: Hold a normal unscripted conversation on camera without dropping into the wrong language
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
Diaspora by Greg Egan - Book
A 1997 novel set in the 30th century, after most of humanity has uploaded itself into "polises," software cities run by and for digital minds, following one such mind who is created from a random mutation and immediately starts wondering what she is. Egan is a working physicist and mathematician, and the novel is considered one of the most rigorous fictional attempts ever made at imagining what life, identity, and consciousness would look like for beings that exist purely as code.
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.
