In partnership with
Two days ago, Dario Amodei asked the whole AI industry to slow down. By Tuesday, Meta's Mark Zuckerberg and Nvidia's Jensen Huang had both told him no thanks, in public, on the record. Underneath that fight is a simpler pattern you'll see all through today's issue: everyone's safety position lines up neatly with their own bottom line. Here are the 7 stories that actually matter for your business today.
The Best Marketing Angle You’ve Never Tried
Everyone wants a marketing hack that makes growing their business easier.
But according to business guru @MarketingMax (9 figures in attributable revenue delivered to clients), one of the best no-cost marketing hacks is finding the one angle that makes people foam at the mouth wanting what you sell.
Lucky for you @MarketingMax put together a quick 2-page Google Doc breaking down how to find yours in 10 minutes, with 20+ real examples you can steal from.
Everyone's AI Safety Argument Matches Their Bank Account
What happened: Over the weekend, Anthropic CEO Dario Amodei published an essay urging the AI industry to slow how fast it builds more powerful models. By Tuesday, Meta's Mark Zuckerberg had rejected the idea, telling reporters that market forces already push labs toward safety because "people won't want to use agents that are misaligned with them," and pointing out that Meta held back its own Muse agent for months without asking anyone else to do the same, according to Fortune. The same day, at Salesforce's Dreamforce conference, Nvidia's Jensen Huang said "safety is an engineering problem, not a legal one" and "we don't need any new laws," calling the whole speed-versus-safety tradeoff "a false choice," per TechCrunch.
Why it matters to you: Look at who's saying what. Nvidia sells the chips a slowdown would leave idle. Meta is racing to catch up at the frontier, and a pause would freeze it in place. Anthropic is asking for restraint while it courts investors ahead of a possible IPO. Nobody in this argument is arguing against their own interest, and that's worth remembering every time an AI company tells you what's "responsible."
What to do about it: When any AI company, big or small, makes a safety claim, ask what that company gets paid to believe.
Apple's New Siri Runs on Google's Brain, Not Apple's
What happened: Apple released the new Siri in beta on September 14th, adding personal context across your messages, mail and photos, on-screen awareness, and tools like Visual Intelligence and Writing Tools, according to Apple's own announcement. The catch: Apple says the features run on "the next generation of Apple Foundation Models, custom-built in collaboration with Google and its Gemini models," meaning the company that spent two years promising a smarter, more personal Siri built it on a rival's AI. It launches in English first, with French, Japanese, Korean, Portuguese and Spanish due next month, and it won't ship initially in the EU or China.
Why it matters to you: If you or your team lean on Siri for anything beyond timers and texts, expect it to get noticeably more useful this fall, on newer iPhones, iPads and Macs only. But also notice the pattern: even Apple, with all its money, decided building its own frontier AI model from scratch wasn't worth the fight. That tells you something about how hard and expensive this technology really is to build.
Most events hand you a binder and a head full of ideas you never get around to implementing. This one hands you a finished asset. Epic Marketing Summit is 2 days in Miami this January where you build a working AI-powered sales funnel on-site, and vibe code your own AI app while you're there. Not a framework to go execute later. The actual thing, done, before you fly home. It's built for owners who are tired of running their sales process on personal time and energy and want it working as a system instead.
I'll be straight with you. I wasn't there last year and I had real FOMO about it, because I know a lot of the trainers and they weren't up there talking theory, they'd actually done the work. It sold out last year, and it's back with a much bigger AI focus for 2027. I'm going in January. If you're serious about building leverage into how you sell, I'd like to see you there. Grab your VIP spot here.
Real People Are Reading Your ChatGPT Conversations
What happened: A 404 Media investigation found that OpenAI employs contractors, under an internal effort called "Project Lily," who read and rate real user conversations with ChatGPT to help improve its responses. One contractor told the outlet, "I don't think they would imagine some contractor somewhere is analyzing the conversations." OpenAI acknowledged it tries to strip personal information before conversations reach reviewers, but confirmed sensitive details can still slip through.
Why it matters to you: ChatGPT has more than 900 million users, and most of them, like most of your customers and employees, probably assume their chats stay private. If you or your staff use ChatGPT for anything involving client details, financial numbers, or anything you wouldn't want a stranger reading, this is your reminder that a human being might actually see it.
What to do about it: Treat any AI chatbot the way you'd treat a work email. Assume someone else could eventually read it, and keep the truly sensitive stuff out.
Meta Just Put a Price Tag on Its AI Features
What happened: Meta introduced "Meta One" on September 15th, a new paid tier across Facebook, Instagram and WhatsApp that unlocks extra AI image and video generation, custom fonts, advanced creator analytics, and a "Meta Business Agent" for handling customer messages, according to Meta's own announcement. Pricing ranges from $2.99 a month for a single feature up to $499 a month for business and creator bundles, while the core free experience stays free.
Why it matters to you: If you run a small business page on Facebook or Instagram, some AI tools you've been using for free may soon come with a monthly bill attached, especially anything related to customer messaging or advanced analytics. Worth checking your Meta Business Suite settings before a feature you rely on quietly moves behind a paywall.
Salesforce Built Its Own AI So It Doesn't Need Yours
What happened: Salesforce and Nvidia announced Koa on September 15th, a reasoning model trained specifically for CRM work like qualifying leads and resolving support cases, built by adapting Nvidia's open Nemotron 3 Super model with synthetic data drawn from Salesforce's own 27 years of CRM experience, according to Salesforce's announcement. The company says Koa matches or beats leading AI models on CRM tasks with three times fewer errors on its internal benchmark, and it runs entirely inside Salesforce's own systems rather than sending customer data out to an outside AI provider. Pilots are underway with Formula 1, UChicago Medicine, Xero and 1-800Accountant, with wider availability planned for winter.
Why it matters to you: If you use Salesforce, this is a preview of what's coming to your CRM: an AI that handles routine lead and case work without your customer data leaving Salesforce's walls. It's also a sign of where bigger software companies are headed. Instead of just plugging in ChatGPT, they're increasingly building their own narrower AI trained on their own data.
The most valuable thing I've gotten out of a mastermind wasn't a tactic. It was the room. I've been a member of Ignite for over 3 years, and it's been a vital part of how I've grown my business. Some of my closest friends came out of that room, and so did a few real business partners. That's the part that never makes it onto a sales page.
The rest of it holds up too. Ignite bundles 6 to 8 live 3-day workshops a year, tickets to 2 in-person multi-day events, weekly mentorship calls with Perry Belcher and his team, and an all-in-one marketing software suite that replaces over $5,000 a month in separate tools. Those workshops normally run $97 to $997 each on their own. Members also get my personal bonus, monthly accountability and business development calls with me. It's built for entrepreneurs still under $1 million a year who want structure and a room of peers to get there faster. If you're looking for a mastermind to move you to your next milestone, this is where I'd point you.
Google's New AI Can Talk and Think at the Same Time
What happened: Google launched two new voice AI models this week: Gemini 3.8 Live, built for speed and scale, and Gemini 3.8 Live Extended Thinking, built for harder, multi-step tasks, according to Google's announcement. Both can detect and switch between 97 languages mid-conversation, keep working on tasks quietly in the background while still talking to you, and narrate progress out loud with lines like "let me check that." They're rolling out now through the Gemini API, Google AI Studio, Search Live, and for paying subscribers, Gmail and Workspace.
Why it matters to you: This is aimed squarely at things like customer support phone lines and voice assistants. If a competitor starts using a voice AI that sounds less robotic and handles more of a call on its own, your customers will notice the difference. That said, on a tough banking-style benchmark, even Google's best version still got the job right barely a third of the time, so don't expect a truly reliable AI phone rep just yet.
A ChatGPT Co-Creator Built an AI That Can't Talk
What happened: Diogo Almeida, who helped build the research behind ChatGPT while at OpenAI, launched a company called TypeSafe AI and a new kind of AI system called Jev, according to TypeSafe's announcement. Instead of generating text, Jev only picks from a preset list of answers with a confidence score attached, which the company says means it can't hallucinate, and it responds in 70 to 500 milliseconds, which TypeSafe claims is 40 to 200 times faster than typical chatbot AI. Pricing is $0.042 per million input tokens with output free, a small fraction of typical chatbot pricing.
Why it matters to you: This isn't something you'd chat with. It's meant to sit quietly inside other software, making small judgment calls like sorting a support ticket or flagging a suspicious order, far cheaper and faster than routing that decision through a full chatbot. If your business software starts feeling snappier at these small background decisions over the next year, this is the kind of technology behind it.
The Bottom Line
Every story this week points at the same thing. Big companies dress up self-interest as principle, whether that's "we don't need regulation" from the company selling the chips, or "trust us to police ourselves" from the company racing to catch up. None of that means AI is useless for your business. It means you should judge these tools by what they actually do for you, not by what the press release says about how responsible the company building them claims to be. Stay skeptical, keep testing what actually helps, and let everyone else ride the hype cycle without you.
Enjoying the Ride?
If today's issue saved you from reading seven other newsletters, forward it to one other business owner who's trying to make sense of all this. And if someone forwarded this to you, welcome, you can subscribe to get tomorrow's issue straight to your inbox.
Talk tomorrow,
Mark Shilensky
Follow me personally on Social Media:
Facebook: https://www.facebook.com/MarkShilenskyPage
Instagram: https://www.instagram.com/markashilensky/
Threads: https://www.threads.com/@markashilensky
LinkedIn: https://www.linkedin.com/in/shilensky/
X: https://x.com/markshilensky
TikTok: https://www.tiktok.com/@mark.shilensky
BlueSky: https://bsky.app/profile/markshilensky.bsky.social