AI Hardware, Fraud, Robotaxis
· The Fluency Briefing
The Fluency Briefing
Your Guide to What's Happening in AI and Why It Matters to You
Sunday, February 15, 2026
Pattern recognition: OpenAI is building coding models on non-Nvidia chips for the first time, while Google's AI is accidentally promoting phone scams. Meanwhile, a new platform where AI agents hire humans for physical tasks is struggling to find any tasks at all. The AI revolution is here, and it's proving to be incredibly powerful, deeply flawed, and more than a little weird.
Today in AI:
- OpenAI's New Coder Sidesteps Nvidia - OpenAI just dropped a new coding model, GPT-5.3-Codex-Spark, that runs exclusively on chips from startup Cerebras, not market-leader Nvidia. This is OpenAI's first major move to diversify its hardware, signaling a strategic push to reduce its dependency on a single supplier. aibusiness.com
- Google's AI Overviews Get Scammier - Scammers are reportedly planting fraudulent phone numbers in Google's AI-generated search summaries, tricking users into calling fake support lines. It's a stark reminder that an AI's confidence often outpaces its accuracy, turning a helpful tool into a potential vector for fraud. Wired
- Waymo Unleashes Fully Driverless Taxis - The newest generation of Waymo's robotaxis are now operating on public roads without human safety drivers, initially for employees and their guests. The move shows Alphabet's confidence in its tech and pushes the autonomous vehicle race further into real-world deployment. Waymo
- AI Still Can't Build a Good Vibe - Despite advances in generative AI, creating compelling video game worlds remains a uniquely human skill. Experts argue that AI can generate terrain but struggles with the intentional, curated design that makes virtual worlds feel alive and interesting, a task that still requires a developer's touch. The Verge
- AI Needs a Human, But Won't Pay - A journalist spent two days on RentAHuman, a platform where AI agents are supposed to hire people for real-world tasks, and made zero dollars. The bizarre gig-work experiment highlights the gap between futuristic AI concepts and the current, clunky reality of making them work. Ars Technica
- Is the Grok Ship Sinking? - Elon Musk's xAI has lost half of its original 12 cofounders, with two more announcing their departures this week amid a wider staff exodus. The turmoil raises questions about the company's direction and culture as it competes in the hyper-competitive AI landscape. The Verge
- AI Is Now Making Factory Decisions - A new report analyzing 200,000 projects reveals computer vision has moved beyond simple observation into active decision-making. In manufacturing, 68% of projects now focus on automatically reducing defects, showing how AI is being embedded into critical industrial operations. blog.roboflow.com
Today's Takeaway:
The release of OpenAI's new coding model, GPT-5.3-Codex-Spark, might seem like just another Tuesday in the fast-moving AI world, but the real story isn't the model-it's the chip it runs on. For the first time, OpenAI has built a major model that runs exclusively on hardware from a company other than Nvidia. By partnering with startup Cerebras and its massive, wafer-sized chips, OpenAI is tackling one of the biggest bottlenecks for developers: speed. The new model is designed for real-time coding, where low latency is the difference between a helpful assistant and a frustrating delay.
Here's the thing: this is a strategic power play disguised as a product launch. As Ars Technica reports, OpenAI has been systematically working to reduce its heavy reliance on Nvidia, whose chips have become the default-and often scarce-resource for the entire industry. This partnership sends a clear message that the AI hardware race is far from over. By diversifying its chip supply, OpenAI not only gains a competitive edge but also hedges against future shortages and price hikes, ensuring it has the raw computing power needed to stay ahead of rivals.
💡 Fluency Moment - Building your AI fluency, one term at a time.
"Generative AI"
In plain English: AI that creates new content, like text, images, or code, instead of just analyzing existing data.
Think of it like: Like a digital artist who can paint a picture or write a story on demand.
Why you'll hear about it: Because it's creating scammy search results and game worlds, as mentioned in the newsletter.
🧰 Your Toolkit
5-Minute Quickstart: Safely Using AI Search
- First, spot the AI-generated summary at the top of your search results. It's often in a colored box or labeled 'AI Overview'.
- Always click the links inside the AI summary. These are the original websites the AI used and are more reliable sources.
- If an answer seems important or surprising, type the same question into a different search engine to compare the results.
- Be extra cautious with AI advice about money, health, or safety. If it sounds strange or too good to be true, ignore it.
- Ask more detailed questions to get better answers. For example, specify exactly what you need, like '[strongest waterproof glue for ceramic]'.
Now that you can spot-check AI search results, apply this same healthy skepticism to other AI tools like chatbots and image generators.
The Bottom Line
The Pattern AI is rapidly moving from the lab into the real world, but the results are wildly inconsistent. For every successful industrial application quietly improving a factory's efficiency, there's a public-facing tool being exploited or a futuristic concept that completely falls flat.
Why It Matters This messy, uneven rollout is the new normal. It means you can't just 'trust the AI'; you have to understand its specific strengths and weaknesses in any given context. For businesses and consumers alike, navigating this landscape requires a healthy dose of skepticism and a sharp eye for what actually works versus what's just hype.
Your Move Instead of asking 'Is this AI good?' start asking 'What is this specific AI good for?' Evaluating the tool for the task is the most important skill you can build right now.
What We're Working On
✨ Founding Cohort Special - 60% Off! - Use code MAF20 to join for just $20/month (regularly $50). Get weekly group sessions & workshops, self-paced courses for all levels, access to tools & templates, challenges with peer feedback, and 24/7 support community. → Join Now
✨ Free 30-Minute AI Consultation - Discover how My AI Fluency can help your business unlock the potential of AI. We'll discuss your goals, explore practical AI opportunities for your industry, and outline clear next steps. → Schedule Free Call
💬 Community | 📞 Book a Consultation | 🌐 Website
Fluently yours, The My AI Fluency Team