Meta’s “Mind-Reading” AI... Rewriting How We Consume Content
Good morning AI entrepreneurs & enthusiasts,
Meta's FAIR research division has developed an AI system capable of anticipating neural responses to video content before viewers even hit play — accomplishing this feat without requiring any brain imaging whatsoever.
Through their TRIBE model, Meta is advancing the boundaries of neuroscience research — while simultaneously crafting what could become a blueprint for content engineered to capture attention at the neurological level.
In today's AI news:
Meta's AI predicts brain responses to videos
OpenAI's reasoner snags gold at programming olympiad
Korean researchers' AI designs cancer drugs
Today's Top Tools & Quick News
META FAIR 🧠 Meta's AI predicts brain responses to videos
News: Meta's FAIR team just introduced TRIBE, a 1B-parameter neural network — also known as a "Trimodal Brain Encoder" — that predicts how human brains respond to movies by analyzing video, audio, and text together, winning first place in the Algonauts 2025 brain modeling competition.
Details:
TRIBE integrates these three sensory channels via a transformer architecture to forecast which of the viewer's brain regions will activate — without requiring any brain scans — thanks to prior training on fMRI data from subjects who watched 80 hours of TV and movies.
The AI correctly predicted over half of brain activity patterns across 1,000 regions, performing best in "association cortices" where sight, sound, and language converge, and outperforming single-sense models by ~30%.
Meta's system also excelled in frontal brain regions that control attention, decision-making, and emotional responses to content.
Why it matters: TRIBE offers researchers an unprecedented tool for modeling and visualizing brain responses to media, deepening understanding of perception and emotion — but also providing a playbook for neural-level engagement design that could make doomscrolling even harder to resist.
GITHUB 🖥️ GitHub CEO Thomas Dohmke to step down
News: Thomas Dohmke, GitHub CEO, will step down at the end of 2025 to start a new venture, citing his “startup roots” as motivation. After his exit, GitHub will not appoint a new CEO—leadership will instead fold into Microsoft’s CoreAI division under Jay Parikh.
Details:
GitHub leadership, including Julia Liuson and Mario Rodriguez, will now report directly to CoreAI.
Under Dohmke, GitHub launched Copilot and doubled AI-related projects.
No replacement CEO planned; integration reflects Microsoft’s broader AI strategy.
CoreAI develops AI tools and platforms, led by former Meta exec Jay Parikh.
Why it matters: This move further entrenches GitHub in Microsoft’s AI infrastructure, likely speeding up AI-powered development tool innovation. Developers and open-source contributors could see shifts in platform priorities, project diversity, and competitive dynamics as GitHub becomes even more integrated into Microsoft’s long-term AI ambitions. Thomas grew GitHub into a household name in the developer community, driving its valuation near $20 billion — and now he's setting out to do it all again. You can take the man out of a startup, but you can’t take the startup out of the man.
AI & DRUG DISCOVERY 💊 Korean researchers' AI designs cancer drugs
News: Researchers at the Korea Advanced Institute of Science & Technology (KAIST) have developed BInD, a groundbreaking diffusion model that can design optimal cancer drug candidates entirely from scratch using only the 3D structure of a target protein — without any prior molecular data or training examples.
Details:
Unlike traditional methods or older AI models, BInD designs both the drug molecule and its binding interaction with the target protein in a single integrated step, considering the binding process from the start to produce stable, effective candidates.
It created molecules that selectively bind to cancer-causing protein mutations (such as specific EGFR variants) while sparing healthy proteins.
BInD simultaneously optimizes binding affinity, safety, stability, manufacturability, solubility, and structural realism.
It reuses successful binding patterns discovered in earlier runs to improve future designs without starting over, integrating real-world chemical rules to ensure physical plausibility.
Why it matters: Drug discovery is one of the biggest beneficiaries of AI acceleration, and BInD represents a leap toward AI-generated precision medicines. By skipping the trial-and-error screening of vast molecular libraries, it could drastically reduce costs and timelines — bringing us closer to a flood of humanity-altering therapies designed entirely by advanced AI.
🛠️ Today's Top Tools
📺 Hera - Create animated videos like CTAs, charts, and ads with text prompts in seconds
🧊 Copilot 3D - Turn images into 3D models with one click
⚙️ Kombai - AI agent for frontend development
📰 Quick News
Chinese AI lab Z AI released GLM-4.5V, a new open-source visual reasoning model that achieves top scores on over 40 different benchmarks.
The U.S. government is reportedly preparing a new agreement with chipmakers Nvidia and AMD that would provide a 15% cut of chip sales to China.
Alibaba announced that its Qwen3 models have been upgraded with ultra-long context capabilities of up to 1M tokens.
Anthropic unveiled new memory capabilities in Claude for Max, Team, and Enterprise users (excluding the Pro tier), enabling it to reference previous chats.
👋🏽 Storyteller, Certified Bubble Developer and Nocode Wizard!
1dThat TRIBE model is wild! Predicting brain activity before you even see the content is almost sci-fi. Big implications for how we make and consume media... and for what data we leave behind. The pace of AI news is starting to feel like a full-time job just to keep up 😅
Professional development coach for recent graduates and young professionals. Get your first job. Develop the skills you need to quickly advance your career in International Relations.
1dThe TRIBE model is both fascinating and a stark reminder of the speed at which AI is evolving. This technology is no longer something of the future; its implications are here, affecting how we interact with content, from entertainment to advertising. The fact that AI can predict neural responses to a video without the need for brain scans is a huge leap forward. It's crucial that we all keep up with these developments. Not only to understand how tools can benefit research or content creation, but also to be conscious consumers. How the technology behind screens can influence our attention, our decisions, and our emotions? This is a call to action for everyone: we cannot be mere spectators. It is essential that we educate ourselves about AI so we can participate in the conversation about its impact and shape a future where it serves humanity ethically and transparently. This case is a clear example of the duality of AI. On the one hand, it opens new doors for neuroscience research. On the other, it gives us insight into how platforms can design content to keep us engaged. This underscores the importance of continually educating ourselves about AI so we can take advantage of its benefits while protecting ourselves from its risks.