In partnership with
Welcome to GPTCheats AI Recap 👋
A quick weekly roundup of the AI news and updates actually worth knowing about.
AI news has been moving fast enough lately that it's easy to see ten headlines in a week and remember none of them.
So I'm trying something a little different with these weekly editions.
Instead of throwing every AI update at you, I'm picking a few stories that I think are actually worth knowing about, especially the ones that tell us something about where AI is heading.
In today's issue:
🧮 AI reportedly solves a decades-old math problem
💸 AI models are getting dramatically cheaper
🚨 1,000+ AI insiders call for better guardrails
🎙️ AI voice cloning makes family scams much harder to spot
Four stories, four very different angles, but all of them say something interesting about where we're headed.
And at the end, I've also included Prompt of the Week — a useful prompt you can try out right away.
Let's get into it.
1,000+ Proven ChatGPT Prompts That Help You Work 10X Faster
ChatGPT is insanely powerful.
But most people waste 90% of its potential by using it like Google.
These 1,000+ proven ChatGPT prompts fix that and help you work 10X faster.
Sign up for Superhuman AI and get:
1,000+ ready-to-use prompts to solve problems in minutes instead of hours—tested & used by 1M+ professionals
Superhuman AI newsletter (3 min daily) so you keep learning new AI tools & tutorials to stay ahead in your career—the prompts are just the beginning
🧮 AI just solved a math problem humans couldn't crack for decades
A new (still-unreleased) AI model reportedly disproved a famous open math problem tied to legendary mathematician Paul Erdős, one about how many points can sit exactly one unit apart on a plane. A top mathematician who evaluated the proof said he'd recommend it for publication in a leading journal.
This isn't just AI answering questions anymore, it's AI making genuine new discoveries in one of the most rigorous fields that exists.
Why it matters: this is a real shift from "AI helps you work faster" to "AI does original research," which is a much bigger deal than it might sound.
💸 AI just got a lot cheaper across the board
OpenAI slashed the price of one of its models by 80% this week, and DeepSeek released a new budget model at a fraction of the cost of the big names, all while ChatGPT quietly crossed roughly 1 billion weekly users.
Companies are increasingly offering a "cheap, everyday" AI option alongside a pricier "smart, deep-thinking" one, so users can pick based on the task.
Why it matters: whatever AI tool you use, there's a good chance it's about to get cheaper or offer a lower-cost option — worth checking your app's settings.
🚨 Over 1,000 AI insiders just sent a warning — but not the kind you'd expect
More than 1,000 employees from OpenAI, Anthropic, Google, and other top AI labs signed a joint statement this week asking the U.S. government for help "pacing" AI progress.
Notably, it wasn't a demand to pause AI or shut anything down — it was a request for better tools and coordination to keep development responsible as things speed up. What's unusual is who signed it: senior scientists and co-founders, not outside critics.
Why it matters: when the people building the technology are the ones asking for guardrails, it's worth paying attention.
🎙️ A real scam warning: AI can now clone a voice well enough to fool your family
A news investigation this week showed just how convincingly AI can now generate realistic voice in different languages, realistic enough to power "it's me, I need help" scam calls that sound just like a loved one.
Police are recommending families agree on a simple, private code word to verify it's really them in an emergency call.
Why it matters: this is a genuinely useful, 2-minute thing to do with your family this weekend, not a hypothetical risk anymore.
⭐ Prompt of the Week
Turn Any Messy Idea Into a Clear Action Plan
"Act as my strategic thinking partner. I’m going to give you a messy idea, problem, or goal, and your job is to turn it into something I can actually act on.
My idea/problem: [IDEA OR PROBLEM]
Desired outcome: [WHAT I WANT TO ACHIEVE]
Current situation: [CONTEXT]
Constraints: [TIME / BUDGET / RESOURCES / OTHER LIMITATIONS]
First, identify what I'm actually trying to accomplish and separate the important information from the noise. Then challenge my assumptions, point out anything I'm overlooking, and identify the biggest obstacles standing between me and the desired outcome. After that, create a practical step-by-step plan ordered by impact, starting with the smallest action I can take today. For each step, explain what needs to be done, why it matters, and what a successful result looks like. If you don't have enough information to make a strong recommendation, ask me the most important questions before building the plan."
Why it works: Instead of asking ChatGPT for generic advice, you're giving it a framework for turning uncertainty into a concrete plan. The final instruction also prevents it from confidently filling in important gaps with assumptions.
That’s it for this week’s Friday Recap.
Hope you enjoyed the new format and found at least one update worth knowing about.
Enjoy your weekend, and I’ll see you Monday with some new prompts
Catch you next issue,
Founder, GPTCheats
