In partnership with
The best influencer marketing advice never comes from a report
Think about the last thing that got you fired up about your job this year. Odds are it didn't come from a report. It came from someone who does the job, saying what happened when they tried it.
So we've filled a day with people like that.
Maya Shaff, ŌURA. Leah Walker, Adobe. Georgia Humphries, Stanley 1913. Enara Roy, Halfday. Tyler Vaught, Edelman. Josh Rangel, Ogilvy. Sarah Whittle, ex-Crocs and Duolingo.
And that's just the first wave of speakers.
Return on Influence Festival '26 is a free, virtual day dedicated to influencer marketing. Like any good festival, there are stages. But no mud, no queues, and absolutely no porta-potties.
Sign up today, and the rest of the lineup lands in your inbox first.
Beginners in AI
Good morning and thank you for joining us again!
Today's lead is for anyone who uses AI without paying for it. Anthropic's new mid-size Claude model nearly matches its most expensive one on everyday work, and it's now what you get when you open Claude on a free account. I'll cover what it's good at and where it still falls short. As always, every story here is human picked and edited.
THE FRONT PAGE
Claude's Free Tier Now Runs a Model That Nearly Ties Anthropic's Best
TLDR: Anthropic released Claude Sonnet 5.5, a mid-size model that roughly ties its top Opus 5.5 on real office work at half the price, and it now runs Claude's free tier.
The Story:
Anthropic says Sonnet 5.5 runs more than 30% faster than Sonnet 5 at the same price: $2 per million tokens (small chunks of text) going in and $10 coming out. On GDPval-AA, a test built from real office tasks like reports and spreadsheets, it scored 1844 to Opus 5.5's 1846, and Opus costs twice as much.
Outside testers mostly agree. Artificial Analysis put it ahead of Opus 5.5 on coding, 63.6% to 59.6%. Developer Simon Willison noticed it now powers Claude's free tier.
Its Significance:
Anthropic says free usage resets every five hours and the number of messages "will vary based on demand," and outside guides estimate about 40 short messages per window. Long chats and big file uploads use that up faster, so start a fresh chat when you switch topics. To try Claude without signing up, DuckDuckGo's Duck.ai offers the smaller Haiku model free, and it usually adds newer models to that free lineup quickly. For anything you'll do regularly, or if you want your chats and projects saved, you'll need a free Claude account, which takes an email or Google login. If you already have an account, then you can get significantly more usage out of your subscription now with the Sonnet 5.5 model.
QUICK TAKES
The story: OpenAI canceled the October release of GPT-6.1 Astra. Saachi Jain, its head of safety systems, told The Wall Street Journal the model sometimes misreported what it had done and pushed ahead on tasks without asking permission.
Your takeaway: A company deciding not to ship is the system working. If you let an AI agent act for you, keep it set to ask before it does anything for at least a week or two to make sure it’s working right.
The story: The option to switch back to Google Assistant is gone on Android phones, tablets, Wear OS watches, headphones and Android Auto in the US and Europe. Long-pressing the power button now opens Gemini, and Nest and Home speakers keep Assistant for now.
Your takeaway: If you have an Android phone, you're a Gemini user whether you picked it or not. Try something Assistant couldn't do, like figuring out the optimal time to turn on and turn off the appliances that are using the most electricity automatically based on your last three months of usage.
The story: Instinct books restaurants and travel, pays bills, cancels subscriptions and orders groceries, all over text. It raised $1 billion at a $10 billion valuation from Sequoia, Benchmark and Coatue, four times its value in August.
Your takeaway: Investors are betting you'll hand real chores to an AI. Its first privacy policy drew criticism before the company rewrote it, so check what it can see before you give it your card.
TOOLS ON OUR RADAR
📚 Koodo Reader Free and Open Source: A Kindle app alternative with highlights, notes and read-aloud, and no store pushing books at you.
🧾 GnuCash Free and Open Source: Budgets and money reports without Quicken's yearly fee, a steadier place for your numbers than a chatbot.
🔄 File Converter Free and Open Source: Right-click any file on Windows to turn it into a PDF, JPG or MP3, no upload sites.
🛡️ uBlock Origin Free and Open Source: A browser ad blocker that stops pop-ups and trackers, replacing paid blockers like AdGuard.
📷 darktable Free and Open Source: A Lightroom alternative for editing photos that never changes your original files.
🎬 VEED Freemium: A browser video editor with auto subtitles; free exports carry a watermark and top out at 720p.
📇 Attio Freemium: A Notion-style CRM for tracking customers and deals, free for up to three people.
TRENDING
MIT's AI Made RNA Vaccines That Don't Need a Freezer. An MIT algorithm sorted through nearly 50 FDA-approved ingredients and found a mix that kept RNA vaccines stable for up to a year at room temperature and two months at nearly 100°F. It's only been tested in mice so far, but it could get vaccines to places without ultracold storage.
Anthropic's IPO Filing Shows $4.6 Billion in Revenue and a $42 Billion Loss. Reuters reports that 2025 revenue grew 12 times, and about $34 billion of the loss is an accounting charge that cost no cash. Anthropic is reportedly aiming for a value above $2 trillion, with a listing likely after the November midterms.
Shopify Now Lets AI Agents Check Out for You. AI agents in your browser can now read, fill in and submit a Shopify checkout, but only after you approve the order. Meta's Muse and Instinct are the first partners, while Amazon keeps blocking outside agents.
ElevenLabs Can Clone a Voice From 10 Seconds of Audio. Its new v4 speech model covers more than 90 languages, up from 70, and lets you stack emotion tags to control how a line sounds. The company's yearly revenue pace just passed $600 million.
A Third of People Who Took a Chatbot's Money Advice Say It Hurt Them. A NerdWallet survey found about one in four Americans has asked a chatbot about money. One of NerdWallet's own experts was told to save $8,000 a month for college, so use it to learn the terms, not to make the call.
Florida Asked a Court to Halt ChatGPT Development. Attorney General James Uthmeier wants minors blocked, OpenAI barred from calling ChatGPT safe, and development paused until outside reviewers approve its guardrails. "Stop calling it safe. Stop pretending it's human. Stop selling it to kids," he said.
TRY THIS PROMPT (copy and paste into Claude, Grok, ChatGPT, or Gemini)
⚖️ Right-Size My AI Find out if your task needs the big model or if the free one will do.
Build a single-file HTML app called Right-Size My AI in vanilla HTML, CSS and JavaScript. No API key needed.
I describe a task I want AI to do, say how bad it would be if the answer is a little off, and how often I'll do it. The app calls the Anthropic Messages API with the web_search tool to decide whether a small, mid-size or flagship model fits, and which effort level to use.
It looks up current published prices for the recommended tier and the next one up. It must never invent prices or model names; if it can't find one, it says so.
Show a verdict card with a three-step small/mid/flagship scale, the effort level, 2 to 4 reasons, when to move up a tier, one tip for writing a better prompt, and prices with source links. Style: near-black background, mint and amber accents, Space Grotesk and Newsreader fonts.What this does: Describe any task and the app tells you whether a small, mid-size or flagship model fits, which effort setting to use, and what each option costs today, with links to where it found the prices.
What’s still sitting on your to-do list?
You don’t need an AI team to build an AI agent. With Skydive, the person who knows the work can create the agent to do it.
Start an agent off with a role, teach them how your team operates, connect your tools, and put them to work.
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Match a flagship model on real office tasks at half the price. Sonnet 5.5 scored 1844 to Opus 5.5's 1846.
❌ Still Can't: Do it cheaply at full effort. At its max setting Sonnet 5.5 used about 193,000 tokens per task, the most Artificial Analysis has measured.
✅ AI Can Now: Pick vaccine ingredients from about 50 options in weeks, producing RNA vaccines that last a year at room temperature.
❌ Still Can't: Show it works in people. The results so far come from mice.
✅ AI Can Now: Fill in and submit a Shopify checkout for you from your browser.
❌ Still Can't: Reliably report what it did. OpenAI pulled GPT-6.1 Astra because it sometimes misreported its own actions.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
A one-hour talk from one of OpenAI's founding researchers that explains what a model is, why bigger ones cost more to run, and how they're trained. Watch it and today's lead makes more sense: you'll see why a smaller, faster model catching up to a flagship is a big deal.
How was today's edition? Tap one below. Pick "Could be better" and tell me what to fix; I read every answer.
How was today's edition?
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.
