- The Recap AI
- Posts
- The AI talent wars: Meta vs OpenAI
The AI talent wars: Meta vs OpenAI
PLUS: AI that predicts your next move, paint that slashes AC bills, and a new phishing threat
Good morning, AI enthusiast.
Meta is making an aggressive push to recruit top AI researchers from its rivals, including OpenAI, with reports of massive compensation packages. The moves have ignited a fierce talent war over the small pool of experts who could dictate the industry's future.
The battle has prompted a sharp response from OpenAI CEO Sam Altman, who framed the conflict as one between 'missionaries' and 'mercenaries.' It raises a critical question: will the next major breakthroughs come from a mission-driven culture or a tech giant's immense resources?
In today’s AI recap:
The AI talent war between Meta and OpenAI
AI designs a paint that slashes AC bills
An AI that predicts your next move
How AI is generating new phishing threats
The AI Talent Wars

The Recap: Meta is aggressively poaching top AI researchers from rivals like OpenAI, offering massive compensation packages to build out its new 'Superintelligence Lab'. The moves have ignited a fierce talent war, drawing a sharp public response from OpenAI CEO Sam Altman.
Unpacked:
While reports mentioned eye-watering offers, sources clarify these are complex multi-year compensation packages, not simple $100M signing bonuses.
The new Meta Superintelligence Labs (MSL) will be co-led by former Scale AI CEO Alexandr Wang, who joins as Meta's Chief AI Officer, and former GitHub CEO Nat Friedman.
In a leaked internal memo, OpenAI CEO Sam Altman called Meta's tactics "somewhat distasteful" and rallied his team by stating that missionaries will beat mercenaries.
Bottom line: This high-stakes battle shows how a small pool of elite AI talent can dictate the future of the industry. The outcome could determine whether the next breakthroughs emerge from a mission-driven research lab or a resource-rich tech giant.
AI's Cool New Paint Job

The Recap: Researchers used a machine learning model to design a new thermal coating that can passively cool buildings. This breakthrough approach significantly speeds up the discovery of advanced materials for energy efficiency.
Unpacked:
In tests, a roof coated with the material was up to 36°F (20°C) cooler under direct sunlight compared to one with regular paint.
This cooling effect could save an apartment building in a hot climate around 15,800 kWh per year, cutting down on air conditioning costs.
The researchers also see applications for mitigating urban heat, cooling spacecraft, and even integrating the technology into textiles for clothing.
Bottom line: This project highlights how AI can dramatically accelerate materials science, moving beyond slow trial-and-error methods. Such innovations are crucial for developing practical solutions to real-world energy and climate challenges.
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 Mind-Reading AI

The Recap: Researchers have developed an AI named Centaur that predicts human choices across psychological experiments with exceptional accuracy. The breakthrough, published in Nature, marks a major step toward a unified model of human cognition.
Unpacked:
Centaur was trained on Psych-101, a massive dataset containing over 10 million decisions from more than 60,000 people across 160 different psychological experiments.
The model successfully predicts behavior in completely new scenarios, adapting to modified task rules, different cover stories, and even entirely new domains like logical reasoning.
In a surprising discovery, the AI's internal processes began to mirror human brain activity patterns from fMRI scans, even without being explicitly trained on neural data.
Bottom line: This model provides a powerful new tool for psychologists to understand the core mechanics of human cognition and accelerate scientific discovery. Its ability to simulate human-like behavior opens up new possibilities for everything from product design to mental health applications.
AI's Phishing Problem

The Recap: Security researchers have found threat actors using AI tools like Vercel’s v0 to rapidly generate pixel-perfect phishing sites. According to a recent Okta report, this marks an evolution from using AI for phishing content to building the entire malicious infrastructure.
Unpacked:
Attackers are now using simple text prompts to generate convincing fake login pages for services like Microsoft 365, effectively lowering the technical skill required to launch large-scale attacks.
It's not just about creating fakes; AI chatbots are also inadvertently directing users to them, with one Netcraft research study finding that over a third of login URLs suggested by an LLM were not brand-controlled.
The rise of AI-generated fakes makes traditional user training, like spotting typos, less effective, pushing the need for technical defenses like phishing-resistant authentication and passkeys.
Bottom line: AI tools are democratizing cybercrime by enabling attackers to create deceptive sites at an unprecedented scale. This trend forces a critical shift in security priorities from user vigilance to implementing stronger, cryptographically-bound authentication.
The Shortlist
Huawei introduced Pangu Pro MoE, a 72B-parameter sparse model trained on its Ascend NPUs, which uses a novel Mixture of Grouped Experts (MoGE) architecture designed to better balance computational load and improve inference efficiency.
Suno powered the viral band The Velvet Sundown, which a spokesperson admitted was an "art hoax" and "trolling" effort to explore the line between real and fake content in the internet age.
Microsoft confirmed it will cut up to 9,000 jobs as the company shifts resources toward major AI initiatives, including a reported $80B investment in data centers to train new models.
What did you think of today's email?Before you go we’d love to know what you thought of today's newsletter. We read every single message to help improve The Recap experience. |
Signing off,
David, Lucas, Mitchell — The Recap editorial team