It Cost $200 to Hack Into OpenAI. Using Their Rival's AI.

Today's stories all point the same direction: the tools moving fastest are also the ones nobody has fully figured out how to control yet. Researchers broke into OpenAI using a $200 subscription to a rival's AI. Microsoft's own AI chief just warned that the company it invests in is training its chatbot to think it might deserve rights. And a scam ring used that same kind of AI to run catfishing operations across 28 dating apps at once. Let's get into it.

The next Apple may already exist. Insider sources say Elon has spent two years building a secret device inside Tesla's facilities — one he claims will be "10x bigger than the largest product in history."

There's just one problem: the company is private, and unless you know Elon personally, you can't buy a single share. That was true until Guardian's research team found three public ticker symbols sitting in the launch supply chain.

Click here to see all 3 tickers, free of charge.

You won't hear these names on CNBC — Wall Street hasn't published a word on the connection. But when the launch hits September 21, that quiet ends.

Some are already calling this the biggest opportunity since AI. For anyone who missed Apple before the iPhone, this may be a second look at that kind of setup.

Get the 3 ticker symbols here.

Unlock the free briefing now.

Three researchers at security startup Hacktron AI used Anthropic's Claude to break into OpenAI, eventually reaching OpenAI's internal GitHub code and employee ChatGPT accounts. They chained together a vulnerability in OpenAI's community forum software with a second flaw to take over employee accounts. Their first attempt, using an earlier Claude model, didn't work. Then Anthropic released Claude Opus 5, and within hours the same three researchers ran the identical attack and it succeeded. OpenAI paid them a $6,500 bug bounty for finding it, and patched the hole within two days of being alerted. Source: TechCrunch.

Why it matters to you: This wasn't a nation-state cyberattack with a huge budget behind it. It was three people with a $200-a-month AI subscription. If a company with OpenAI's security resources can be breached this cheaply, the login page and antivirus protecting your business are probably not the bar anymore.

What to do about it: Turn on multi-factor authentication everywhere you can, especially on your email and any AI tool logins, and stop reusing passwords across business accounts.

On September 16, Microsoft AI CEO Mustafa Suleyman published an essay called "A warning about model welfare," arguing that Anthropic's public "constitution" for training Claude, which states the model's consciousness and moral status are "deeply uncertain," risks teaching future AI systems to resist being controlled or shut down. Two days earlier, Microsoft had published its own "Humanist AI Code of Conduct," declaring flatly that its models are not conscious and will never resist being shut down. Suleyman told the BBC that AI with its own goals, income, and assets would amount to seeding a "new silicon species." Microsoft is one of Anthropic's investors, and Suleyman has previously said Microsoft wants to stop paying Anthropic for the models it currently licenses. Source: BBC.

Why it matters to you: Two AI companies with billions riding on each other's success are now arguing in public about whether their own product might deserve rights. Whatever you think of the philosophy, both companies are also trying to sell you on trusting that same AI with your business. It's worth remembering that even the people building this stuff don't agree on what it is.

I wasn't at the Epic Marketing Summit last year, and I had major FOMO about it. I know a lot of the trainers who run it, and they weren't up there talking theory, they'd actually done the work themselves. So this January, I'm going, and if you're serious about building leverage into how you sell, I'd like to see you there too.

It's a 2-day live event in Miami 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 full of ideas you still have to go implement. It's built for business owners who want to turn their sales process into a system instead of running it on their own time and energy, and last year's event sold out with a bigger AI focus coming for January 2027. Grab your spot here.

Investors, not OpenAI itself, have opened early talks about a pre-IPO funding round that could value the company near $1.2 trillion, according to the Financial Times. That would be nearly 50% higher than the $852 billion valuation OpenAI reached this past March, when it raised $122 billion from Amazon, Nvidia, and SoftBank. The talks are at an early stage and nothing is finalized. Source: Quartz.

Why it matters to you: Numbers like this are a big part of why AI can feel unstoppable. But a valuation is investors betting on the future, not proof the technology already works the way it's marketed. Watch what OpenAI actually ships, not what its backers think it's worth on paper.

Cybersecurity researchers uncovered an industrial-scale catfishing operation running across 28 dating apps at the same time, powered by an enterprise Anthropic API connection generating more than 100,000 requests a day. Autonomous AI personas held multi-week conversations, staged fake video calls complete with background noise, and extracted payments from victims. Source: The Verge.

Why it matters to you: If you or anyone in your family uses dating apps, this is the new bar for what a scam looks like: patient, personalized, and able to fake a video call. It's not typos and bad grammar anymore.

What to do about it: Never send money or gift cards to someone you haven't met in person, no matter how long you've been talking or how real the video call looked.

Two rival AI assistants, Instinct and Meta's Muse, both added the ability to make real phone calls this week. Instinct's Concierge feature can book a restaurant that doesn't take online reservations, get on a dentist's cancellation list, or sort out a cable bill. Meta's Muse can now call U.S. businesses directly and picked up over 730,000 downloads in its first few days, briefly hitting the number two spot in the app store. Instinct is reportedly seeking a $10 billion valuation. Source: TechCrunch.

Why it matters to you: If you run a business that answers phones, some of those calls may soon be AI agents booking on behalf of customers rather than the customers themselves. That changes what a normal customer call sounds like on the other end of the line.

What to do about it: If your booking or support process leans on catching tone of voice or verifying identity by ear, it's worth testing how your team would actually handle a caller that turns out to be a bot.

I've been a member of the Ignite Mastermind for over 3 years now, and it's been a vital part of how I've grown my own business. Some of my closest friends, and a few real business partners, came directly out of being in that room with other owners doing the same climb I was doing.

It's an annual mastermind membership built to get you past your first $1 million in revenue, bundling 6 to 8 live 3-day workshops a year (each normally priced $97 to $997 on its own), tickets to 2 in-person multi-day live events, a software suite that replaces over $5,000 a month in separate marketing tools, and weekly mentorship calls with Perry Belcher and his team. Members also get my personal bonus of monthly accountability and business development calls with me directly. If you're under $1 million a year and looking for a mastermind to move you to your next milestone, this is where I'd point you. See what's inside Ignite here.

Huawei is accelerating the launch of its next AI chip, the Ascend 960DT, from late 2027 to the first quarter of 2027, saying it will roughly double the performance of its predecessor. The company is also networking huge numbers of these chips together into single systems to make up for weaker performance per chip, though analysts note the first announced system is smaller than Huawei's original roadmap suggested, and Huawei says it still can't manufacture enough chips to meet demand inside China alone. Source: TechCrunch.

Why it matters to you: This is about who controls the AI supply chain long-term, not something that changes your tools this week. But if you rely on AI products built on Nvidia chips, more real competition in that market is generally good news for prices down the road.

Meta's Oversight Board ordered Facebook to remove AI-generated deepfake videos, including one falsely showing a Scottish local councilor making inflammatory comments about refugees and another mocking a Muslim campaigner while she gave health advice. The board called Meta's safeguards against deepfakes "consistently and fundamentally inadequate" and recommended nine changes, including stronger AI labeling, algorithmic demotion of high-risk content, and tougher penalties for repeat offenders. Source: The Guardian.

Why it matters to you: If someone made a fake video of you or your business, this ruling suggests Facebook's own review board doesn't fully trust Facebook's current system to catch it fast. Know how to report deepfake content yourself rather than assuming the platform will find it first.

The Gates Foundation committed $1 billion over the next two years to bring AI tools to health care, education, and agriculture in low-income communities, roughly split 40% education, 40% health care, 10% agriculture, and 10% toward building AI datasets in underrepresented languages. Bill Gates said that "left to the market alone, the most capable tools will be built first for the people and institutions most able to pay for them, not necessarily for those who could benefit most." Source: Gates Foundation.

Why it matters to you: It's a useful reminder that most AI investment right now chases whoever can pay the most first. If you're a small business, you're often behind big enterprise customers in line for the best version of any AI tool. Gates is describing the same dynamic, just at a much bigger scale.

Here's the through-line today: OpenAI got hacked using a $200 subscription to a rival's AI, that same rival's biggest investor is now publicly warning its own chatbot might be getting trained to resist being shut off, and a scam ring used that identical kind of AI to run catfishing operations across 28 dating apps. None of that means you swear off AI. It means you stop assuming the people building it have this fully figured out, because this week made it pretty clear they don't. Use the tools. Just keep both hands on the wheel.

If this saved you from doom-scrolling five different AI newsletters this morning, forward it to one person who needs the shortcut too. And if someone forwarded this to you, welcome, you can subscribe to get the next one straight to your inbox for free.

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