AI is creating new billionaires at a record pace
The AI gold rush isn’t just minting billionaires, it’s breaking records. With 498 AI unicorns worth a combined $2.7 trillion (100 of them born since 2023), the sector is creating wealth faster than any tech wave before it. Mega-rounds for Anthropic, OpenAI, and Anysphere have produced fortunes for founders barely out of their twenties.
Most of that money is still locked up in private equity, but secondary markets, acquisitions, and soaring Bay Area real estate prices show the cash is starting to move. Wealth managers are circling, but history says the AI-rich will first double down on more AI bets before learning the value of diversification.
OpenAI’s cheap models, which you can download
After years of keeping its weights close to the vest, OpenAI just went public, sort of. The company dropped two open-weight models, gpt-oss-120b and gpt-oss-20b, marking its first such release since GPT-2 in 2019.
They’re not fully open source, but they are designed to be cheaper, customizable, and runnable anywhere, from cloud clusters to your laptop. OpenAI teamed with Nvidia, AMD, Cerebras, and Groq to ensure cross-chip compatibility, while running extra safety checks to filter out dangerous data and test for malicious fine-tuning.
You can grab them on Hugging Face or GitHub under Apache 2.0, use them in LM Studio or Ollama, or access them via Amazon, Microsoft, and Baseten. With rivals like Meta, Mistral, and DeepSeek already in the open-weight game, OpenAI’s move is part competition, part outreach, and a reminder that billion-dollar AI models aren’t just for billion-dollar data centers.
AI Joins the Cybersecurity Squad
Generative AI has supercharged cybercriminals, making phishing, deepfakes, and malware slicker and faster than ever. The defense? AI agents of our own. Companies like ReliaQuest and Syniverse are rolling out “agentic” AI that can sift through logs, quarantine phishing emails, restrict compromised accounts, and automate tedious security tasks like verifying executives’ devices on overseas trips. These tools act as tireless teammates, freeing human analysts for higher-level work and addressing the industry’s talent shortage. Adoption is still in the “trust but verify” phase, but with attackers already wielding AI at scale, many see cybersecurity as the place where agentic AI will take root fastest.
Even a $4B Founder Can Get an AI Reality Check
Emma Grede, the powerhouse behind Skims and Good American, admits she was “using AI like a 42-year-old woman” until Mark Cuban gave her a jolt. On her podcast Aspire, Cuban revealed he runs 60 AI apps, sparking Grede to enroll in AI courses at Wharton and Harvard this fall.
While she’d already incentivized her teams to use AI with cash bonuses two years ago, she’d mostly left the experimentation to her younger staff. Now, she’s eyeing AI not for personal productivity, she claims she’s maxed out, but for smarter strategic decisions and big bets.
Cuban’s warning to entrepreneurs who ignore AI: “You’re f–ked… It’s like saying you don’t need the internet.”