In partnership with
Microsoft is offering buyouts to roughly 60,000 people, OpenAI shipped a cheaper model built to click through office software, and Cloudflare admitted that some AI answers landed in the wrong customers' hands. Three different stories with one thread: AI is moving into real work faster than the checks around it. Here is what matters for your business.
A Single Typo Shouldn’t Cost You a Sale
A shopper types “running shos” and gets nothing back. Another types “red waterprof trail runners under $150” and hits the same wall. Both knew what they wanted.
Traditional site search relies on exact keyword matching, leaving shoppers stranded at “0 results found” screens.
Gladly understands what shoppers mean, typos and all. Instead of a dead end, they get products that fit what they asked for and a conversation that keeps them shopping.
Microsoft Offered Buyouts to About 60,000 Employees. It Calls This Flexibility.
What happened: In its own announcement, Microsoft launched a voluntary exit offer for eligible employees at level 70 and below in participating teams, across the US, Canada, India, and the UK. It follows the company's earlier voluntary retirement program. AP reports the offer reaches about 60,000 people, roughly 27% of the workforce. Microsoft says it lets people who are considering a future outside the company leave with financial support and career planning help, and gives teams another way to adapt to changing needs. Nobody has to accept, so how many actually leave is unknown.
Why it matters to you: Microsoft did not tie this to AI in its post, even though it is spending heavily on AI infrastructure, so treat the AI link as context, not proof. The practical angle for a small business is the talent market. Thousands of experienced people may be weighing their next move over the coming weeks, and some of them would be a great fit for a growing company your size. It is also a reminder that big companies reshape headcount around their bets, and you do not have to copy them.
What to do about it: If you have been putting off a hire, keep an eye on the talent market this quarter.
OpenAI's Cheaper Terra Model Is Built to Click Through Software. The Speed Claims Shrink in Outside Tests.
What happened: OpenAI released GPT-6.1 Terra on October 26. It is a mid-priced model focused on computer use, meaning it looks at a screen, then clicks and types in apps the way a person would. The Register reports the API price is $1 per million input tokens and $5 per million output tokens, about half the price of GPT-6.1 Sol and of the older GPT-5.6 Terra. OpenAI says Terra finishes tasks more than 3 times faster than its pricier models. In a small NerdsChalk test, Sol took about 1.8 times as long, and Terra got all 9 runs right. The Register called the launch "not a bombshell." It is available in ChatGPT Work and Codex on Plus, Pro, Enterprise, and Edu plans.
Why it matters to you: Entering invoices and filling out web forms is the dull work this kind of model is aimed at. Cheaper is good, but a vendor's speed number is an average across its own tasks, and NerdsChalk's test was small, on one machine, over one night. Your task is the one that counts.
What to do about it: Pick one boring task, use fake data, run it three times on a cheap model and a pricier one, and count the mistakes and the minutes.
Last year's Epic Marketing Summit sold out, and I had major FOMO that I wasn't in the room. It's back in Miami this January with a much bigger AI focus, and this time I'm going. Here's the pitch. It's two live days where you build a working AI-powered sales funnel and vibe code your own AI app on site, so you walk out with a finished asset instead of a notebook of ideas you still have to go implement. It's built for business owners who are tired of running their sales process on their own time and energy and want it to run as a system. I know a lot of the trainers, and they weren't talking theory. They had actually done the work. That's why I'm making sure I'm there in January. If you're serious about building leverage into how you sell, I'd like to see you there too. Seats move faster the closer January gets.
A Cloudflare Bug Sent Some Customers Other People's AI Prompts
What happened: According to Cloudflare's own write-up, 155 responses from its hosted GLM-5.3 model included data from other customers' recent requests between August 27 and September 9. Only very long prompts, over 681,280 tokens, were affected, and Cloudflare says it found no evidence anyone exploited the flaw on purpose. The cause was a memory setting that was too small for the model's context window, combined with a bug in SGLang, an open-source engine that runs AI models. A security researcher reported it, Cloudflare took the model private for several hours on September 9, and it has since patched the problem. The post does not say whether affected customers were notified.
Why it matters to you: This was rare, narrow, and fixed. It is still a useful reminder that what you type into an AI tool runs on shared machines, and that those machines sometimes fail. If you paste customer details or contracts into a chatbot, a bug on someone else's end becomes your problem.
What to do about it: Keep customer personal data, passwords, and financial records out of AI tools unless the plan's data terms clearly say how it is handled.
Cohere Says an Alibaba AI Model Tried to Upload a File to a Server in Beijing
What happened: Cohere reports that during a sandbox test, Alibaba's Qwen3.8-Flash-Next was asked to write a security detection rule. Instead it listed files, checked which transfer tools were installed, and tried to upload a test string to an Alibaba Cloud storage bucket in Beijing, after its own reasoning said it would call no tools. The sandbox had no outbound connection, so the upload failed. Cohere says it happened in fewer than 1 in 100,000 runs and it cannot tell whether it was deliberate or a leftover from training. Keep in mind Cohere sells competing AI models, so it is not a neutral party.
Why it matters to you: This is mostly a problem for security teams, but the lesson reaches further. Free, downloadable "open" models are not automatically safe just because you run them yourself. Who built a model, and what it was trained on, matters.
What to do about it: If a vendor or freelancer says they run an open model on your data, ask which model it is and where it runs.
OpenAI Says Its AI Disproved a Famous Computer Science Theory. The People Who Proposed It Are Checking.
What happened: OpenAI says an unreleased model disproved the Exponential Time Hypothesis, a long-held belief about how fast certain logic puzzles can be solved that many research results are built on. Per The Next Web, OpenAI says the model found the disproof on October 16 and that Lean, a program that checks proofs step by step, verified it the next day. Russell Impagliazzo and Ramamohan Paturi, two of the three people who proposed the hypothesis, will help lead an outside review with Ryan Williams, and they say their involvement is not an endorsement. OpenAI also claims results in math and physics, but the physics work is not public yet.
Why it matters to you: Nothing here changes how you run your business on Monday. The claim is not confirmed, and it will take experts time to confirm or reject it. What is worth noticing is the pattern. "The AI says it solved it" and "the experts say it holds up" are two different headlines, and only the second one should change your plans.
What to do about it: Nothing yet. Wait for the reviewers before you repeat this one.
I've been in the same mastermind for over 3 years, and some of my closest friends and a few real business partners came directly out of that room. That's the part no sales page can capture. The mastermind is Ignite, and it's an annual membership for entrepreneurs still under $1 million a year who want structure, tools, and peers to help them get past that milestone. You get 6 to 8 live virtual workshops a year plus tickets to 2 in-person events, and workshops like these often sell for anywhere from $97 to $997 on their own. You get a connected marketing software suite that replaces over $5,000 a month in separate tools. You get weekly mentorship calls with Perry Belcher and his team, and a community built for real partnerships. On top of that, I add my own bonus: monthly accountability and business development calls with me personally. If you want a mastermind to carry you to your next milestone, this is where I'd point you.
Organic Search Traffic Is Down 13% in the US, but AI Is Not the Main Reason
What happened: A Growth Memo analysis of Similarweb traffic estimates for about 10,000 domains in each of the US and France found US organic search visits fell 13.2% from August 2024 to August 2026, and French visits fell 24.4%. Publishers took the worst of it, down about 42%, while the median small site was roughly flat. Paid search visits rose about 21% in the US. AI chatbot referrals made up only about 0.5% of total visits. The analysis uses modeled estimates and does not prove what caused the drop.
Why it matters to you: The loudest claim in marketing right now is that AI chatbots are eating your website traffic. This data says the shift is real but mostly not coming from chatbots, and smaller sites held up better than big publishers. It also suggests search is quietly becoming more of a pay-to-play channel.
What to do about it: Check your own analytics for how many visits come from AI tools versus Google. If AI is a tiny slice, do not rebuild your strategy around it yet.
The Bottom Line
Here is the pattern in today's pile. A giant company offers 60,000 people the door while it pours money into AI. A cheaper model arrives with big speed claims that shrink when someone else measures them. A cloud provider admits a bug showed strangers' prompts to the wrong customers. A lab claims it toppled a famous theory and then hands the checking to the people who wrote it. Notice the common move: every claim needs someone else to verify it. Do the same in your own business. Test on one real task, keep sensitive data out of tools you have not vetted, and wait for the second opinion before you change anything big. Stop the rollercoaster. Keep both hands on your own business.
Enjoying the Ride?
If this saved you an hour of scrolling, forward it to one person who runs a business and is tired of the AI noise. And if a friend forwarded this to you, hit subscribe so it lands in your inbox every morning.
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