Do you like shiny new releases? If so, today is your day! 🤩
[In case you missed it: I’m now answering all your questions, every Sunday, Inside the Kitchen]
🤯 MYSTERY AI LINK 🤯
This link leads to one of the most interesting things I’ve seen in AI recently.
🛠️ NEW AI TOOLS 🛠️
💬 PROMPT OF THE DAY:
Build a morning routine that fits [wake time], [available minutes], [energy level], [responsibilities], and [one priority habit]. Give me a minute-by-minute sequence, explain why the order works, and include a five-minute emergency version for rushed mornings.
What’s cookin’? Anthropic just launched Claude Opus 5.5, its new flagship AI model, which provides better performance at a significantly lower cost. Anthropic says Opus 5.5 is especially strong at coding, research, computer use, and long, complicated tasks. One early tester reportedly migrated a 680,000-line codebase in less than a day. The bigger story may be efficiency: Anthropic says typical workloads cost about 40% less than Opus 5, while the model generates responses more than 30% faster. It also scored better on Anthropic’s safety evaluations and is designed to be more resistant to prompt injection and unwanted autonomous actions.
🤔 Hunter’s take: The AI race isn’t just about making models smarter anymore. It’s about making intelligence cheap enough to use everywhere. A model that can handle hours of professional work while costing dramatically less changes the economics of AI agents. That’s when AI moves from “helpful chatbot” to something companies can realistically leave working on entire projects.
Image: OpenAI
What’s cookin’? OpenAI has launched GPT-6 Sol and GPT-6 Luna, two faster, cheaper versions of its newest GPT-6 family. Sol is built for demanding work like coding, research, business workflows, and computer tasks, while Luna is the lower-cost option for high-volume everyday use. The biggest change may be the price: OpenAI cut API costs roughly 50% compared with GPT-5.6, while claiming significant improvements in accuracy, coding, and agent performance. Sol’s input price drops from $4 to $2 per million tokens, while Luna falls to just $0.10. OpenAI also improved prompt caching, which can make long-running AI agents even cheaper.
Fun FryAI Fact: “Prompt caching” is like a fry cook remembering your usual order instead of rereading the whole ticket every time you come back. If your AI request keeps reusing the same instructions or information, it can reuse those recipes, making the response faster and cheaper.
🤔 Hunter’s take: As you can see, intelligence is getting cheaper. That changes what companies can actually afford to automate. Tasks that were too expensive to hand to an AI agent thousands of times per day suddenly start making financial sense. If today tells us anything, it’s that the AI race is becoming a price war, too.
👊 Real questions. Real answers.
One insider email. Every Sunday.
📈 HOW ARE THE BIG AI PLAYERS DOING?
🤖 HAS AI REACHED SINGULARITY? +1.5%
How close are we to the point where AI surpasses human intelligence across nearly every domain and begins advancing beyond our control or prediction?
What do you think of today's newsletter?
👆 I read all feedback. Thank you!
Hi — I’m Hunter, a PhD candidate whose work has appeared in major academic journals and popular tech outlets. I founded FryAI to make staying ahead of AI clear, accessible, and fun.