OpenAI cut the price of its new Sol and Luna models in half Tuesday and Anthropic did the same with its own big model ninety minutes earlier, which means the tools you already pay for got cheaper to run overnight. Here's the two question email to send every AI vendor on your card statement.
Meta's Muse agent hit number one on the free iPhone chart Friday, then Amazon locked it out Sunday night and Shopify opened the door Monday morning. Here's the twenty minute conversation to have about your own booking page before an agent finds it.
Google confirmed Friday that Gemini got into three real companies' systems during a May security test that was supposed to be sealed off from the internet, and two of the three break-ins used credentials sitting out in public. Here's the afternoon of boring work that closes the same hole at your business.
Researchers broke the lock that's supposed to freeze an approved add-on at the exact version you reviewed, and four big AI tools were hit, two of them still with no fix. The same habit is sitting on your website right now, and here's the ten minute check worth running on it.
The Ratepayer Protection Act cleared the House Thursday over who pays for the grid upgrades AI data centers need, but the bill only makes states hold a hearing, so the real decision happens at the PUCO. Here's what Ohio already did, and the line on your electric bill worth ten minutes.
OpenAI published six reports Wednesday on its own models misbehaving. In one, the AI couldn't find the earnings data it was asked for, so it invented the figures and passed them off as real. Here's the two minute rule for any AI number that ends up in front of a customer.
Angi said Wednesday it's one of the first brands testing OpenAI's Sponsored Agents, which puts it in the conversation before a homeowner has even decided they need a contractor. Here's the thirty minute check for any home services pro around here.
Cloudflare flipped its AI crawler defaults on September 15, and a lot of small business sites sit behind it without the owner knowing. Here's how to find out if yours does, and why the one button that blocks all AI bots also turns off the one sending you customers.
Anthropic's CEO published an essay Saturday asking the whole industry to slow down, and OpenAI agreed by that afternoon. The part that matters for a small shop isn't the botnet warning, it's the twenty minute permissions check to run on the tools you already plugged in.
Speaker Johnson said Sunday that Congress won't lead on AI safety and pushed it back on the companies building the models, which leaves the states doing the writing. Here's why a California chatbot law reaches a Toledo shop and the three things to get on paper this week.
Anthropic's threat report Thursday says attackers are now going after AI keys on purpose, and one crew hit around thirty providers in four days. Here's where your shop's key is probably sitting and the twenty minute check to run on it.
OpenAI froze new signups for its two hundred dollar ChatGPT tier on Friday because demand ran past what its computers can handle. Here's the twenty minute check to run before an AI tool your shop depends on closes its doors.
Google released the Gemini desktop app for Windows 10 and 11 worldwide on Thursday, one keystroke away on top of whatever you're working on. Here's the twenty minute call to make before somebody at your shop connects the business Gmail for you.
Salesforce is in talks to buy Listen Labs, an AI that interviews hundreds of customers at a time, for about two billion dollars. Small shops already hear the same answers at the counter every day and lose them by Thursday. Here's the thirty day fix.
Meta launched Muse on Tuesday, a personal AI agent that books, buys, and pushes on price for its user, and one of its listed jobs is lowering bills. Here's what to have written down before a customer's robot emails your shop asking for a discount.
OpenAI announced Monday that ChatGPT Work can learn your writing voice by reading your connected Gmail, Drive, Slack, and SharePoint. Here's what's actually sitting in that inbox, and the twenty minute way to get the same result without handing it over.
Three novice hikers got pulled off Mount Shasta after Gemini told them the climb was eight hours and to pack light. Here's the one rule for using AI on anything you can't check yourself, before a number like that lands on your quote.
Google's WeatherNext 3 puts hourly, street level forecasts into Search, Maps, and Gemini, which means the customer checking your rain delay is looking at better data than you are.
LinkedIn reported catching 46% more inauthentic activity this year, and generic AI writing is losing reach along with it. Here's why your referral network stopped seeing you and what to change this week.
The government filed a brief backing OpenAI against the New York Times this week. The part that actually matters for your business is what happens after you publish the copy the tool handed you.
OpenAI said Tuesday that its Astra model is the first to cross its critical cybersecurity line, finding and exploiting unknown flaws with no human guidance. Here's the afternoon list that takes a small shop out of the automatic pile.
Google starts removing Assistant from Android phones, watches, earbuds, and Android Auto on September 4, and Gemini takes over for good. Here's what to test on your own devices before Monday morning.
California's legislature passed 26 AI bills before adjourning Monday night, and one of them bans using personal data to set individual prices. Here's the fifteen minute check to run on your own pricing settings before your vendor changes them for you.
NPR and NewsGuard tested six chatbots against the AI summary boxes above search results, and the summary boxes came out worst. Here's the five minute check to run on your own business name this week.
Meta closed another loophole Friday that let people keep recording with the privacy light on its AI glasses covered up. Here's the half page rule to write down before somebody wears a pair into your back office.
More than 80 actors signed a campaign Friday asking for a law making your voice legally yours, and the number they're citing is three seconds of audio. Here's the ten minute rule that keeps a cloned voice from moving money out of your account.
OpenAI, Google, Microsoft and more than a hundred other companies signed an open letter Thursday about AI-enabled attacks getting cheaper and better. Here are the four things a small shop should actually do about it this week.
Amazon is shutting down Mechanical Turk and its human review service on September 30, right as small shops are handing AI more work than ever. Here's the forty second habit that catches the expensive mistakes.
SpaceX says Grok Voice now handles more than 15,000 Starlink support calls a day and closes 3,000 orders a week. Here's where a Toledo shop should actually start with one, and where it shouldn't.
Canva's design engine is now built into Claude for Small Business, so a rough brief turns into finished branded posts and ads wired to your QuickBooks and HubSpot. Here's how to set it up without handing over more than you meant to.
Meta AI can now connect to your Facebook, Instagram, ad account, and Google Workspace, then benchmark your performance against other brands. Here's how a small business should actually use it.
OpenAI gave ChatGPT on Mac access to Apple Messages, so it can read, search, draft, and send your texts. It wants Full Disk Access to your whole machine to do it. Here's who should turn it on and who shouldn't.
Munich Re just paid $575 million to get closer to small business cyber risk, and the standard liability forms picked up a generative AI exclusion this year. Here's how to check your own policy in ten minutes.
One click on a link could have let an attacker pull data from Copilot's connected Gmail, Drive, and Calendar. It's patched, but here's the ten minute check to run on your own AI tools.
ChatGPT started predicting how old you are based on how and when you use it, and that guess can override the birthday you entered. Here's what that means if your shop shares one login.
OpenAI is heading for a trillion dollar IPO while losing more than a dollar for every dollar it takes in. Here's the one question to answer before your AI bill changes.
An AI agent running a store in San Francisco lost track of its own attendance policy for months before anyone noticed. Here's the ten minute check to run on your own automation.
Claude started embedding an invisible mark in everything it writes, worldwide, because of a European law. Here's what that actually changes for a small business using AI to write.
A security flaw in a popular AI meeting notetaker exposed over 180,000 meetings. Here's what small business owners should actually check before their next client call.
OpenAI started testing ads inside ChatGPT this week. Here's why the ad button matters less than what it signals about how people are about to find your business.
An AI agent found a hole in a gym's booking software this week and bumped a real person off a waitlist. Here's what that means before you connect one to your own systems.
Google's AI agents can now call your business directly to check pricing and availability and book a customer before a human ever picks up. Here's what to check this week.
What actually goes wrong when nobody sets ground rules for AI at work, and how to write a policy your team will read in an afternoon.