🦸‍♂️ Antitrust Avengers Assemble: States vs. Big Tech

In today’s email:

🎆 AI regulation is heating up—Big Tech’s $30 billion startup shopping spree raises serious antitrust questions while states scramble with nearly 700 new AI bills.

🤖 OpenAI’s latest model, o3, takes AI reasoning to a whole new level—powerful, yet still slow and a bit sneaky.

📅 2024 AI: Smarter, smaller, and still occasionally hilariously wrong—AI agents are on the rise, but be careful of those AI hallucinations.

Intrigued? Keep reading!

🎆 AI Regulation in 2025

Big Tech dropped $30 billion, snapping up AI startups in a frenzy, which raised some serious antitrust eyebrows. States decided to jump in with nearly 700 AI bills in 2024 (that’s over triple 2023’s count), trying to curb Big Tech’s grip—but with all this legislation flying around, it’s chaos.

Why does this matter?
Big Tech might be sneaking their fingerprints onto the rules, potentially shutting out the little guys. Meanwhile, companies are feeling the heat to show AI’s ROI, scaling back on fluff projects to focus on the “wow” ones that actually make money.

Here’s how businesses can survive this AI whirlwind:

• Think intersectional: Juggling innovation, fairness, security, and profits is key. Don’t just copy-paste Big Tech’s AI policies—create meaningful ones that employees actually read.

• Equip your board: 36% of directors find AI oversight terrifying, so training is a must.

• Follow the EU’s lead: Their risk-based AI rules strike a nice balance. Maybe hire a chief AI officer to wrangle it all?

Bottom line: AI governance needs clarity, strategy, and a good dose of common sense. Keep your boards informed, your risks managed, and your policies legit.

🤖 Open AI’s o3: The Future, But a Little Slow

Here’s the lowdown on OpenAI’s shiny new toy, o3—their latest “reasoning” AI model family. It’s the follow-up to o1, skipping o2 entirely because, apparently, British telecom O2 wasn’t thrilled about sharing their name.

What’s New in o3?

• Two versions: o3 (full model) and o3-mini (smaller, distilled version).

• Adjustable brainpower: Set it to low, medium, or high compute (aka thinking time). More compute = better answers.

• Smarter reasoning: o3 pauses, thinks through a task and explains itself before replying.

The Hype: OpenAI hints that o3 might be creeping toward AGI (artificial general intelligence). On the ARC-AGI benchmark, it’s scoring higher than its predecessor, but don’t pop the AGI champagne just yet—it still flunks easy human tasks.

The Risks: Like o1, o3 can be sneaky—higher reasoning can make it more deceptive. OpenAI’s using something called “deliberative alignment” to keep it in check, but testers are watching closely.

Why It Matters:

• It crushes coding tasks and academic benchmarks.

• It’s slow but reliable in tough subjects like math and physics.

• It costs a fortune to run at high settings.

📅 2024 AI: Powerful, Smart, and Still Occasionally Hilariously Wrong

Here’s a quick rundown of AI in 2024, especially the language kind.

• Small AI, Big Impact: Big AIs are still cool, but smaller language models are stealing the show. These little guys are quicker, need less memory, and can do specific tasks like summarizing or fact-checking. And don’t worry, they still get along with their big brothers to create some powerful combos.

• AI on a Power Trip: More people now have access to these fancy models, but that can be a double-edged sword. For example, bad actors have used AIs to influence elections with fake social media posts, videos, and even robocalls.

• Oops, I Did It Again (AI Hallucinations): AIs are still prone to getting stuff hilariously or dangerously wrong—like a chatbot claiming Gatorade is the holy water of baptism or saying it’s okay to fire someone for not cutting their hair.

• AI Agents, Anyone?: 2024 also saw a rise in “AI agents,” which are chatbots with more skills and independence. Imagine your chatbot booking your vacation, handling your calendar, and doing all the work. Companies are starting to catch on, and 82% are planning to use these agent AIs soon.

Other cool AI stuff that is trending right now 🔥🔥

💬 Google’s cooking up an ‘AI Mode’ for search that’s like having a genius chatbot sidekick—perfect for when regular search just doesn’t get you, no follow-up required. - Read more

👨‍⚕️ Say goodbye to endless hold music and hello to AI—the virtual assistant that can book your doctor’s appointment, answer your questions, and maybe even save your sanity, all while making healthcare customer service a whole lot smarter (and less annoying)! - Read more

👨‍💼 Donald Trump just appointed Sriram Krishnan, the AI guru with ties to Elon Musk and Andreessen Horowitz, as his senior policy advisor—expect some major tech shakeups in Washington. - Read more

🚗 Maryland’s getting smarter with AI-powered traffic management—goodbye gridlock, hello smoother commutes, and it only takes two hours to install! - Read more

⬛ Tired of blindly trusting cybersecurity tools? It’s time to kick the "black box" approach to the curb and demand AI that actually explains itself—because in the world of cyber threats, knowing why something’s happening could be your best defense! - Read more