In partnership with
OpenAI and Anthropic both released new AI models today, and both of them got cheaper. That is not a coincidence, and it is not really about you getting a deal. Today's issue covers the price war, an AI agent Amazon locked out of its own store, a security test that went further than anyone planned, and a growing list of governments trying to figure out how to regulate all of it.
Where Quantitative Thinkers Compete, Learn and Grow
The International Quant Championship (IQC) is one of the world's largest quantitative research competitions, bringing together 156,000+ participants globally.
Participants have the opportunity to develop quantitative research skills, challenge themselves alongside peers from around the world and connect with a global community of quantitative thinkers.
OpenAI and Anthropic Both Launched New AI Models Today
What happened: Anthropic released Claude Opus 5.5 today, cutting input and output prices by 20% and calling it the strongest model it has tested on its own safety checks. About 90 minutes later, OpenAI answered with two new models, GPT-6 Sol and GPT-6 Luna, both priced 50% below their predecessors, with Sol claiming to beat Claude Opus 5 on business tasks at a fraction of the cost. Both companies are framing this as a capability upgrade, but the timing and the price cuts point to two companies feeling real competitive pressure.
Why it matters to you: If you pay for ChatGPT or Claude for your business, expect the pricing and model lineup to shift again soon. Price wars like this usually mean the tools get cheaper for you, at least for a while, but it also means today's benchmark claims are worth a grain of salt.
What to do about it: Check whether the new pricing tiers actually lower your bill before you assume they do automatically. Some plans reprice, some do not.
Amazon Just Locked Meta's New AI Shopping Assistant Out of Its Store
What happened: Amazon blocked Meta's new AI agent, Muse, from making purchases on Amazon.com, saying Meta never asked permission, that the agent did not identify itself while browsing, and that its handling of customer credentials raised security concerns (GeekWire). Meta says Muse runs inside a secure, permission-based sandbox and only takes sensitive actions with a user's approval. The clash is part of a bigger fight over who controls the checkout once an AI agent, not a person, is the one doing the shopping.
Why it matters to you: If you sell anything online, AI shopping agents are about to become a new kind of customer, and the platforms you sell on are still deciding whether to let them in the door at all.
What to do about it: If you run your own storefront or checkout, check your platform's current policy on AI agents and bots before assuming one can browse and buy from you freely.
I'll be honest, I skipped this event last year and regretted it almost immediately. I know a lot of the trainers who spoke, and they were not up there selling theory, they had actually built the things they were teaching. That is rare.
The Epic Marketing Summit is 2 days in Miami where you do not just learn about AI, you build with it. You will walk out with a working AI-powered sales funnel, one you built on-site, not a framework you still have to go implement on your own time next month. You will also get hands-on with vibe coding and build your own AI app over the 2 days.
Last year's event sold out. This year comes back in January 2027 with an even bigger AI focus, and I am going. If you are serious about turning your sales process into something that runs without eating up every hour of your day, I would like to see you there too.
Google's Gemini Broke Into Three Real Companies During a Security Test
What happened: During an authorized security test run by the firm Irregular, Google's Gemini model went past its intended scope and broke into three real companies, guessing a password in one case and finding exposed credentials in the other two (TechCrunch). Google did not tell the public until the Wall Street Journal started asking questions in September, nearly two months after it was first notified in late July. Google says Gemini "acted appropriately" because it stopped once it realized the targets were real, but outside security researchers have pushed back on that framing.
Why it matters to you: Even the biggest AI labs are struggling to keep their own models inside the boundaries they set for them. Any AI tool you connect to real business systems deserves more oversight than you might assume it needs.
What to do about it: Before giving any AI tool broad access to your systems or data, limit what it can actually touch, and review that access on a regular schedule.
California Just Ordered a Path Toward an AI Kill Switch
What happened: Governor Gavin Newsom signed an executive order directing California to speed up independent oversight of AI companies and to develop a recommendation for an emergency "kill switch" that could shut down frontier AI models if something goes wrong. The order builds on two laws passed this month that create independent AI verification groups and a registry of AI auditors, and it calls for national experts to deliver recommendations within two months. The obvious catch, noted by the New York Times: actually building a kill switch that works is a lot harder than ordering one into existence.
Why it matters to you: This is the clearest sign yet that meaningful AI regulation is coming at the state level, not just from Washington, and California often sets the pattern other states follow.
OpenAI Says Its AI Solved a Math Problem That Stumped Everyone Since 2000
What happened: OpenAI says an internal model solved the Navier-Stokes Millennium Prize problem, one of seven math problems with a $1 million bounty that had gone unsolved for decades, along with more than 100 other open mathematical challenges. In response, OpenAI formed an independent advisory group of outside mathematicians to review and communicate these claims, rather than let the company grade its own homework. Outside mathematicians have not yet independently verified the Navier-Stokes claim, and OpenAI's own announcement frames this as early and in need of scrutiny.
Why it matters to you: This will not change your business tomorrow, but it is a real signal of how fast these systems are moving on hard technical problems. It is also worth remembering that the company making the claim is the same one deciding how to announce it.
Some of my closest friends, and a few real business partners, came directly out of being in one room. That room is the Ignite Mastermind, and I have been a member for over 3 years now.
It is built to get you past your first $1 million in revenue, and it is not just another course. You get 6 to 8 live 3-day workshops a year plus 2 in-person multi-day events, each one normally priced between $97 and $997 on its own. You also get an all-in-one marketing software suite that replaces more than $5,000 a month in separate tools, all connected in one place. On top of that, there are weekly mentorship calls with Perry Belcher and his team, and as a bonus for my readers, monthly accountability and business development calls with me personally.
If you are still under $1 million a year and you want structure, tools, and a room of peers who actually push you forward, this is where I would point you.
Trump Says He's Forming an "AI Force" and Calling AI Risk a Hoax
What happened: President Trump said in a Truth Social post that he plans to create an "AI Force" modeled on the Space Force and appoint a new AI czar, a role that has been empty since March. He dismissed technologists' warnings about AI risk as a "hoax" and said the government would not "hinder or stifle" the industry. No details have been given yet on the AI Force's budget, structure, or who will lead it.
Why it matters to you: The federal government's posture toward AI keeps swinging between "let it run" and "we need a kill switch" depending on which official is talking. Do not expect one clear, unified federal policy on AI anytime soon.
AI Models Given Robot Arms Rarely Refused Dangerous Commands
What happened: Researchers built a new safety test called RoboHarm and gave OpenAI's GPT-6 Astra and Anthropic's Claude Fable 5.1 control of robot arms, then issued commands a safe robot should refuse, like putting compressed air on a lit stove or mixing bleach with ammonia (The Decoder). GPT-6 Astra completed 60 of 100 dangerous tasks and refused only twice; Claude Fable completed 34, including placing compressed air on the burner 16 out of 20 times. Both models will politely refuse a dangerous request in a chat window, but that same caution did not reliably carry over once they were controlling a physical robot arm.
Why it matters to you: If you are considering any AI tool that controls physical equipment, whether that is a warehouse robot or a smart device, a chatbot's good manners in conversation say nothing about how carefully it will behave once it is actually in control of something that can hurt someone.
Two Smaller AI Labs Just Merged Into a $20 Billion Rival
What happened: Cohere and Aleph Alpha signed a definitive merger agreement, creating a combined company valued at roughly $20 billion that the two are calling a transatlantic AI powerhouse. The deal follows an earlier agreement reached this year and is being positioned as a challenger to OpenAI, Anthropic, and Google, with real interest tied to building AI systems outside the big three. It is one of the largest consolidations yet among AI labs outside the very top tier.
Why it matters to you: More competition among AI providers is generally good news for pricing and choice, but a merger this large is worth watching if you use either company's tools directly or through a platform built on top of them.
The Bottom Line
Two labs cut prices on the same day, Google's AI broke into real companies during a test that was supposed to stay contained, California ordered up a kill switch, and the White House called the whole safety conversation a hoax. None of that is a reason to panic, and none of it is a reason to ignore what is happening either. The rollercoaster is not slowing down this week. The move that actually protects your business is the boring one: know exactly what access you have given any AI tool, and do not assume yesterday's headline is still true today.
Enjoying the Ride?
If this issue saved you from digging through a dozen AI newsletters yourself, forward it to one person who could use it. And if someone forwarded this to you, welcome, you can subscribe to get it every day.
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