Good morning, AI enthusiast.

A 60-year-old math problem that has long stumped researchers is now being cracked by AI. New reinforcement learning techniques are making progress where humans have been stuck for decades, potentially opening up new fields of discovery.

The work shifts AI from just solving problems to navigating challenges where solutions may not even exist. Could this be the first step toward building a mathematical 'crystal ball' for predicting and preparing for complex, real-world events?

In today’s AI recap:

• AI's new mathematical foresight

• GitHub's CEO departs as platform merges with Microsoft

• The future of coding jobs amid AI boom

• 8 trending AI Tools

AI's Mathematical Crystal Ball

The Recap: Researchers are using reinforcement learning to tackle a 60-year-old math problem, making progress where humans have been stuck for decades. This mathematical breakthrough could lay the groundwork for AI that can predict major crises like stock market crashes or climate disasters years in advance.

Unpacked:

• The system uses two AI agents working together: a "player" attempts to solve the problem with basic moves, while an "observer" learns to bundle these into powerful "supermoves" to make bigger leaps.

• This approach successfully solved or simplified large sets of potential counterexamples to the conjecture that had been open for over 30 years, with the results being independently confirmed by a separate study.

• The ultimate goal is to create tools that find rare but critical patterns in enormous datasets, with potential applications in cryptography, medicine, and finance.

Bottom line: This work shifts AI from just finding optimal solutions to known problems to navigating challenges where solutions may not be known to exist. By training on abstract mathematics, these models could one day give us the foresight to prepare for complex, real-world events.

Learn from this investor’s $100m mistake

In 2010, a Grammy-winning artist passed on investing $200K in an emerging real estate disruptor. That stake could be worth $100+ million today.

One year later, another real estate disruptor, Zillow, went public. This time, everyday investors had regrets, missing pre-IPO gains.

Now, a new real estate innovator, Pacaso – founded by a former Zillow exec – is disrupting a $1.3T market. And unlike the others, you can invest in Pacaso as a private company.

Pacaso’s co-ownership model has generated $1B+ in luxury home sales and service fees, earned $110M+ in gross profits to date, and received backing from the same VCs behind Uber, Venmo, and eBay. They even reserved the Nasdaq ticker PCSO.

Invest for $2.90/Share

_{Paid advertisement for Pacaso’s Regulation A offering. Read the offering circular at invest.pacaso.com. Reserving a ticker symbol is not a guarantee that the company will go public. Listing on the NASDAQ is subject to approvals.}

AI Tools of the Day

1. 🎬 Pika - Bring your wildest ideas to life by generating and editing captivating videos from simple text prompts or images.

2. 🎨 LoraAI - Train custom AI models to replicate any character, object, or artistic style for truly unique image generation.

3. 🤖 Alethea AI - Create, own, and monetize interactive AI characters on the blockchain as unique, intelligent NFTs (iNFTs).

4. ⚙️ ragie.ai - Empower your LLM applications with a powerful RAG-as-a-Service API to securely interact with your own private data.

5. 💻 CodeGeeX - Supercharge your IDE with a powerful AI coding assistant that generates, translates, and explains code across multiple languages.

6. 🛡️ Troj AI - Secure your machine learning models against adversarial attacks and hidden vulnerabilities with an automated AI penetration testing platform.

7. ↔️ Codia AI - Instantly convert Figma designs, screenshots, or even prompts into production-ready frontend code for web and mobile apps.

8. 🎶 Riffusion - Generate novel music by visualizing it as a spectrogram and using stable diffusion to compose unique audio from text prompts.

Explore the Best AI Tools Directory to find tools that will 10x your output 📈

GitHub's Leader Departs

The Recap: GitHub CEO Thomas Dohmke has announced his resignation, and the developer platform will be integrated more tightly into Microsoft's CoreAI division. This move signals a significant shift from the operational independence GitHub has maintained since its acquisition in 2018.

Unpacked:

• Microsoft is not replacing the CEO, instead having GitHub’s leadership report directly to its CoreAI team.

• The move aligns GitHub with CoreAI leader Jay Parikh’s vision to build an "AI agent factory" for enterprise customers.

• Dohmke is leaving to become a startup founder again but will stay through the end of 2025 to guide the transition.

Bottom line: This change solidifies GitHub’s role as the central engine for Microsoft's AI-first developer strategy. The platform's future innovations will now be directly shaped by Microsoft's broader AI ambitions, impacting its massive user base of over 150 million developers.

AI Training

I Used Lovable + n8n to Build an AI Web Developer Agent (free template)

The Recap: In this video, I’ll show you how to a fully-automated web developer AI agent inside of n8n that can research websites, plan out a re-design by writing Product Requirements Document (PRD), and can generate a website all on its own using Airtop + Lovable.

P.S. We also launched a free AI Automation Community for those looking to build and sell AI Automations — Come join us!

The Coding Job Jitters

The Recap: A Purdue computer science grad's viral story about her only interview offer coming from Chipotle is spotlighting a tough job market for new developers, sparking widespread debate about AI's role in the future of coding.

Unpacked:

• Recent data highlights the trend, with unemployment for recent computer science and computer engineering graduates hitting 6.1% and 7.5% respectively—more than double the rates for many humanities degrees.

• The hiring process itself is transforming, with companies using AI to screen resumes and creating an "AI loop" where automated applications are met with automated rejections, often within minutes.

• Meanwhile, many experienced developers see AI as a productivity tool for boilerplate code and debugging, not a replacement for the complex problem-solving and domain knowledge that define their roles.

Bottom line: The long-standing promise that a computer science degree is a golden ticket to a stable career is being rewritten. Future success will likely depend less on rote coding ability and more on a professional's capacity to leverage AI for efficiency while focusing on creativity, strategic thinking, and complex system design.

Where AI Experts Share Their Best Work

Join our Free AI Automation Community

Join our FREE community AI Automation Mastery — where entrepreneurs, AI builders, and AI agency owners share templates, solve problems together, and learn from each other's wins (and mistakes).

What makes our community different:

• Real peer support from people building actual AI businesses

• Complete access to download our automation library of battle-tested n8n templates

• Collaborate and problem-solve with AI experts when you get stuck

Dive into our course materials, collaborate with experienced builders, and turn automation challenges into shared wins. Join here (completely free).

The Shortlist

SK Hynix projects the AI memory market will grow 30% annually through 2030, driven by immense demand for high-bandwidth memory (HBM) chips.

Apple will integrate OpenAI’s new GPT-5 model into its upcoming iOS 26 and macOS Tahoe releases, upgrading the optional ChatGPT query feature.

Meta named conservative activist Robby Starbuck an AI bias advisor as part of a lawsuit settlement, after its AI chatbot falsely linked him to the January 6th Capitol riot.

The Boston Public Library is partnering with OpenAI and Harvard Law School to use AI to digitize and enhance searchability for its vast archive of historical government documents.

Signing off,

David, Lucas, Mitchell — The Recap editorial team