- Nextool AI
- Posts
- Claude May Have Solved Math’s Biggest Mystery
Claude May Have Solved Math’s Biggest Mystery
Plus: Signal Just Drew a Line Against Apple and Google's AI
Three major AI developments reveal how quickly things are changing. Anthropic's Claude may have helped solve a 167-year-old mathematical mystery. Meanwhile, Signal is blocking AI assistants from accessing private conversations. And new research from METR reveals something surprising about AI progress. Models are getting better at complex tasks but still struggle with judgment. Together, these stories raise an interesting question: How powerful is AI becoming, and where should we draw the line?
In today’s post:
AI May Have Solved Math's Biggest Mystery
Your Private Messages Have a New Problem
AI Is Getting Smarter. But There's a Catch.
SPONSORED BY
It's time to set up your holiday streaming campaign
Are you ready to reach holiday shoppers on CTV? For many streamers, the season is already underway.
Last year, according to Roku internal data, holiday content searches on Roku saw an 86% increase during October. And in November, those searches rise an average of 39% weekly. This surging holiday spirit is mirrored by consumer activity, as about half of Roku streamers shopped Black Friday and Cyber Monday deals in 2025.
Meeting those shoppers is easier than you might think. Begin with the seasonal outcome you hope to drive: awareness, online sales, leads, store visits, or another measurable goal. From there, define your audience, upload your creative, set the budget and campaign dates, and establish measurement before launching. You can even create a shoppable ad unit that lets streamers buy with a click of their remote.
Our holiday playbook offers a month-by-month guide to reaching shoppers during this all-important time of year.
First-time spenders get a $5,000 match on the first $5,000 spent with code BEEHIIV5K. Terms apply.
What’s Trending Today
RESEARCH
A 167-Year-Old Math Problem. And Claude May Have Cracked It

Image credits: BBC
Anthropic just made a claim that could change mathematics. On November 20, its mathematician announced a potential breakthrough. With Claude's help, he claims to have solved the Riemann hypothesis. A problem that has defeated mathematicians for 167 years. But there's one important catch: the proof remains unverified.
Here's why this matters.
The Riemann hypothesis has remained unsolved since 1859. Proposed by Bernhard Riemann, it concerns the distribution of prime numbers. Its solution would represent a historic mathematical achievement.
Anthropic mathematician Levent Alpöge says Claude helped find a proof. He publicly shared his claim on November 20, 2026. It came just one day after Anthropic's IPO.
This wasn't a simple chatbot answer. The work involved mathematical reasoning and formal proof software. Alpöge released millions of lines of code using Lean.
The potential breakthrough reaches beyond mathematics. Prime numbers underpin important areas of number theory. Understanding their distribution could deepen several branches of mathematical research.
The proof still needs independent verification. Mathematicians must examine its logic and confirm its conclusions. Until then, the Riemann hypothesis cannot be considered definitively solved.
Claude had already demonstrated progress on related mathematics. Earlier research improved results connected to the Riemann hypothesis. This latest claim represents a much more ambitious step.
The bigger story is how scientific discovery might change. AI could become more than a tool for calculations. It might help researchers discover ideas they couldn't reach alone.
And that last point deserves attention. For years, AI has been praised for answering questions. Now we're watching it potentially help solve unanswered ones. That's a very different kind of capability. The interesting question isn't whether AI replaces mathematicians. It's whether humans and AI can discover more together. If this proof survives scrutiny, it could mark a turning point. Not because machines suddenly understand everything. But because our tools may help us reach previously unreachable ideas. And even if this particular proof fails, the ambition matters.
POLICY
Signal Just Exposed AI's Biggest Privacy Problem

image credits: Signal
Signal is drawing a line against AI-powered assistants. On November 20, Signal announced new privacy protections. The reason? AI assistants can now read what's on your screen. Even messages protected by end-to-end encryption. And that raises a question few people are asking.
Here's what you need to know.
Signal is blocking AI access to private conversations. New protections restrict screenshots on iOS 27. Android protections also target AI assistants attempting screen capture.
Apple and Google are changing how phones work. Their AI assistants can analyze screen content to perform tasks. But that convenience introduces new privacy risks.
End-to-end encryption isn't the whole story anymore. Signal protects messages during transmission. But operating-system AI could capture content after it's decrypted.
Microsoft's Recall already raised similar concerns. Signal previously blocked Recall screenshots on Windows 11. Now the same privacy battle is reaching smartphones.
Users can disable Signal's screenshot protections. But doing so could expose conversations to AI systems. That includes messages belonging to other participants.
Signal says its protections remain incomplete. Blocking screenshots doesn't prevent every possible access method. AI agents might still retrieve information through device files.
Signal wants Apple, Google, and Microsoft to act. Developers need reliable ways to exclude apps from AI access. They also need clearer documentation explaining how these systems operate.
The bigger issue goes beyond Signal. AI assistants are becoming part of our operating systems. And that changes who controls our personal information. Privacy used to mean controlling what apps could collect. Now we must consider what our operating systems can see. AI assistants promise convenience, and that convenience has real value. But useful technology shouldn't quietly weaken existing privacy protections. Signal's response highlights a problem the entire industry must solve.
GROWTH
AI Can Solve Complex Problems. So Why Does It Still Struggle With Judgment?

image credits: Metr
AI models are improving faster than many expected. A new report from METR reveals just how quickly. Some AI capabilities are still growing at an exponential rate. But the findings expose a surprising weakness. AI can complete difficult tasks without consistently making good decisions.
Here's what the research reveals.
AI capabilities are still improving rapidly. METR's original research found that AI task horizons doubled roughly every four months. Its latest report suggests this progress continues.
AI performs especially well when success is measurable. Coding and mathematics have clear answers. That makes these tasks easier to evaluate and improve through training.
Judgment remains a major weakness. METR tested AI on more open-ended assignments. Models struggled when success required interpretation rather than verification.
AI couldn't design useful research evaluations. Researchers asked models to create new AI tests. Not one evaluation met METR's standard for research usefulness.
Even philosophical reasoning exposed limitations. Models were asked to assess philosophical arguments. Their judgments fell short of human performance in the study.
The latest models are showing signs of improvement. Anthropic's Opus 5.5 achieved borderline performance on evaluation design tasks. A year earlier, models were clearly below the required standard.
The biggest unanswered question concerns AI research itself. Can AI eventually choose promising research directions independently? Researchers still lack enough evidence to predict when that happens.
And there's another complication. Measuring judgment is much harder than measuring coding ability. Even the researchers acknowledge limitations in their evaluation methods. We're getting very good at building AI that produces answers. But intelligence isn't only about finding the right answer. It's also about knowing which questions deserve attention. That requires judgment, context, and an understanding of uncertainty. AI is improving at these skills, but evidence remains limited. The next breakthrough might not be a smarter coding assistant. It might be an AI that knows what matters.
Free Guides
My Free Guides to Download:
🚀 Founders & AI Builders, Listen up!
If you’ve built an AI tool, here’s an opportunity to gain serious visibility.
Nextool AI is a leading tools aggregator that offers:
500k+ page views and a rapidly growing audience.
Exposure to developers, entrepreneurs, and tech enthusiasts actively searching for innovative tools.
A spot in a curated list of cutting-edge AI tools, trusted by the community.
Increased traffic, users, and brand recognition for your tool.
Take the next step to grow your tool’s reach and impact.
That's a wrap:Please let us know how was this newsletter: |
Reach 150,000+ READERS:
Expand your reach and boost your brand’s visibility!
Partner with Nextool AI to showcase your product or service to 140,000+ engaged subscribers, including entrepreneurs, tech enthusiasts, developers, and industry leaders.
Ready to make an impact? Visit our sponsorship website to explore sponsorship opportunities and learn more!

