Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Today we're introducing TRIBE v2 (Trimodal Brain Encoder), a foundation model trained to predict how the human brain responds to almost any sight or sound. Building on our Algonauts 2025 award-winning architecture, TRIBE v2 draws on 500+ hours of fMRI recordings from 700+ people to create a digital twin... show more

AI at Meta

876,232 subscribers

6,982,503 views • 7 months ago •via X (Twitter)

Education Health & Wellness Science & Technology

Anya Rossi• Live Now

Private livecam show

32 Comments

Dhiran7 months ago

500 hrs of brain scans to predict what makes us click... zuck's building the ultimate engagement algo from inside our heads. wild times

AI at Meta7 months ago

Without any retraining, TRIBE v2 can reliably predict the brain responses of individuals it has never seen before, achieving a nearly 2-3x improvement over previous methods for both movies and audiobooks We’re releasing the model, codebase, paper, and demo to help researchers advance neuroscience, apply brain insights to build better AI, and use computational simulation to speed up breakthroughs in neurological disease diagnosis and treatment. 🔗 Paper: 🔗 Model: 🔗 Code:

Raúl R Romero7 months ago

Normally such technology would excite me, but Meta just lost a court case where they DELIBERATELY addicted children and ruined their mental health for profit.

shaun7 months ago

we can sell up to 80% of a user's visual field without inducing seizures

Brooklyn7 months ago

Claude: upload my brain to Facebook. Make no mistakes.

✮ Hood Satoshi7 months ago

Meta just dropped a model that predicts how your BRAIN reacts to video content. So I built an app with it. Paste any YouTube, TikTok, or IG link and get: → Predicted attention curves over your entire video → Emotional intensity mapping → Cognitive load analysis → A live brain visualization that lights up as your video plays → Exact timestamps where viewers lose focus It's basically VidIQ on steroids — except instead of optimizing for the algorithm, you're optimizing for the human brain. Built in a few hours using TRIBE v2. This tech is insane.

Kraken7 months ago

Can we use it to simulate the feeling of 10x-ing our bags?

Connor Burke7 months ago

So this means ad agencies and brands (Coca-Cola, Nike, etc.) can test hundreds of ad versions or product trailers in minutes without focus groups. They see which ones trigger the strongest brain responses for attention, emotion, or "want-to-buy" feelings saving that focus group cost. Algorithms now can be trained to get the user scrolling more for additional ads as well if Facebook can see how our brain reacts to a post… Interesting development!

Aman7 months ago

Impressive. I can think of a dozen different ways to use this right now, like training a social media algorithm to maximize neural activity in the ventral striatum keeping users scrolling reels for hours. Please never do this😭

Flipcoder 🚀7 months ago

Someone at meta right now

Sigmaking69697 months ago

How the FUCK is this not a crime yet? Meta pays “internet safety advocacy groups” 2 BILLION dollars to get age verification to collect even MORE data about us then they announce THIS?! WHAT THE FUCK.

M R X7 months ago

advertisers will license this within two years. a model that predicts what makes your brain react to sight and sound. Meta built the perfect tool for the wrong hands.

Mike7 months ago

So we’re at the “analyze your brain for ads” stage now

Devin7 months ago

Literally can’t keep up

igmut7 months ago

specifically because it's coming from meta: creepy af

aryan 🧠⚡️7 months ago

Congratulations to all the incredible scientists at Reality Labs, spending their time optimizing Reels engagement…

Linus ✦ Ekenstam7 months ago

Sick. is this powering the wrist band?

MacroDog7 months ago

You clowns will realize no one wants this and the market will punish your stock until you cut the shit.

Trader Insights Media7 months ago

Meta just announced TRIBE v2, an AI that can predict exactly how YOUR brain responds to anything you see or hear. TRIBE v2: trained on 500+ hours of brain scans from 700+ people. It predicts neural activity across 1,000 brain regions in real-time. Why this is massive for $META as an investment: ✔️Facebook/Instagram/Reels → content optimized for neural engagement, not just clicks ✔️Ray-Ban Meta glasses → biometric + brain-response data at scale ✔️AI assistants → trained to respond how your brain actually prefers ✔️Healthcare play → brain twins for neurology, mental health diagnosis ✔️BCI future → foundation model for brain-computer interfaces Elon has Neuralink. Apple has neural interface patents. Meta has TRIBE v2. @grok connect the dots for investors. Meta has TRIBE v2 (neural prediction), Ray-Ban glasses (real-world data), Llama (language AI), V-JEPA 2 (video AI), and 3.2 billion daily active users generating behavioral data. Is Meta building a closed-loop system for mapping, predicting, and influencing human cognition at scale? #Meta #TRIBEv2 #Neural #AI #BrainActivity #BREAKING #TechNews

paul7 months ago

These amazing technological advancements are underscored by the fact that they will almost certainly just be used to show us more ads

Nabeal Khan7 months ago

“yeah bro i choose my own ad budgets because i’m smarter than meta”

RegenFlow AI7 months ago

Meta really said “let’s build an AI that knows exactly how your brain’s gonna light up to any pic or sound” and then casually dropped that it was trained on 700 real human brains for 500+ hours 😭 Bro went full sci-fi villain just so Reels can hijack your dopamine harder than ever.

Aaryan Kakad7 months ago

i went through the code, paper and their demo properly. dropped a complete overview of TRIBE v2 in very simple words. check it out here:

Kristen Ruby7 months ago

I know of only one person who wanted to do this exact exercise. His name was J E.

SMA 🏴‍☠️7 months ago

kinda scary

Daniel7 months ago

Folks, never trust the Zucc - he's literally planning to monetize ur brain with 3D ads. Wtf

helix7 months ago

This is how they will control us. If the brain can send a signal, it means they can send one back 🤷‍♂️

sdmat7 months ago

Wait, you spent millions training a model that predicts generic fmri output? Why? Which use cases does this serve that the training data did not? It is pointless for medical application as the whole purpose there is to detect atypical behavior. What does this accomplish?

scoopy trooples7 months ago

nope!

Ēṭēr̥ṇāl̥7 months ago

This is cool on paper but kinda creepy in practice. Meta training on 700 brains to predict reactions sounds like the ultimate ad targeting upgrade, not pure science. We've seen how even basic eye tracking hooks attention for hours. Zero shot on new people? Wild. How long till this quietly shapes what content even gets shown to us?

𝓓𝖗. 𝓥𝖔𝕏 𝓞𝖈𝖚𝖑𝖎 ᴹᴰ 👁️‍🗨️7 months ago

Thought crime soon

B7 months ago

This is prediction, not access. The model learns statistical mappings from stimuli → averaged brain responses (fMRI signals). That’s impressive, but it’s not “reading minds” or capturing individual thoughts. A few key constraints • fMRI is slow and coarse (seconds, millimeters) • you’re predicting group-level patterns, not unique internal states • “zero-shot” here means generalization within the same data structure not new physics So yeah, solid progress for neuroscience and modeling. But it’s still correlation + inference, not direct measurement of subjective experience. @grok

Related Videos

Meta's new AI can predict your brain better than a brain scan. TRIBE v2 is a foundation model trained on 1,000+ hours of brain imaging data from 720 people. You feed it a video, sound clip, or text, and it predicts: > Which brain regions light up > How strongly > And in what order When tested on people it had never seen, the model's predictions were actually more accurate than most real brain scans (which get distorted by heartbeats, breathing, and movement). Researchers then replicated decades of classic neuroscience experiments entirely inside the software. No scanner, no human subjects. The model correctly identified the brain's face recognition center, language network, and emotional processing regions on its own. But look at what else Meta has been building: > Ray-Ban smart glasses that see and hear what you do > A wristband that reads nerve signals > And now a model that predicts how your brain responds to any piece of content There's no evidence these are all connected... but Meta now has a complete picture of attention, from the stimulus to the neural response.

Meta's new AI can predict your brain better than a brain scan. TRIBE v2 is a foundation model trained on 1,000+ hours of brain imaging data from 720 people. You feed it a video, sound clip, or text, and it predicts: > Which brain regions light up > How strongly > And in what order When tested on people it had never seen, the model's predictions were actually more accurate than most real brain scans (which get distorted by heartbeats, breathing, and movement). Researchers then replicated decades of classic neuroscience experiments entirely inside the software. No scanner, no human subjects. The model correctly identified the brain's face recognition center, language network, and emotional processing regions on its own. But look at what else Meta has been building: > Ray-Ban smart glasses that see and hear what you do > A wristband that reads nerve signals > And now a model that predicts how your brain responds to any piece of content There's no evidence these are all connected... but Meta now has a complete picture of attention, from the stimulus to the neural response.

Rowan Cheung

29,939 views • 6 months ago

The future of marketing is prediction on a neural level The demon was trained on the brains of 4 people, to predict how their brains would react to video stimuli. And finally, we beat meta’s tribe v2 on its benchmarks. And that exact model, now is available to you TODAY. Psywar will be the leader for VSLs in every domain. Never forget that. A VSL isn’t a VSL unless it’s a psywar VSL

The future of marketing is prediction on a neural level The demon was trained on the brains of 4 people, to predict how their brains would react to video stimuli. And finally, we beat meta’s tribe v2 on its benchmarks. And that exact model, now is available to you TODAY. Psywar will be the leader for VSLs in every domain. Never forget that. A VSL isn’t a VSL unless it’s a psywar VSL

Krma

46,471 views • 1 month ago

Starting today you can try our new foundation research model for audio generation. The demo includes Zero shot TTS, Text to sound effects, Infilling and more! Try Audiobox ➡️

Starting today you can try our new foundation research model for audio generation. The demo includes Zero shot TTS, Text to sound effects, Infilling and more! Try Audiobox ➡️

AI at Meta

515,818 views • 2 years ago

Introducing Soniox TTS v2, our most powerful text-to-speech model yet. Soniox TTS v2 brings extraordinary voice quality, expressive control through audio tags, exceptional precision, high-fidelity voice cloning, more than 60 languages, natural language mixing, and low-latency streaming together in one model. Built from the ground up with a new model architecture, audio codec, and highly optimized inference engine, Soniox TTS v2 delivers frontier TTS at just $0.70 per generated hour. Available globally today as tts-rt-v2. Try it out: Turn the sound on and hear what’s possible. 🔊

Introducing Soniox TTS v2, our most powerful text-to-speech model yet. Soniox TTS v2 brings extraordinary voice quality, expressive control through audio tags, exceptional precision, high-fidelity voice cloning, more than 60 languages, natural language mixing, and low-latency streaming together in one model. Built from the ground up with a new model architecture, audio codec, and highly optimized inference engine, Soniox TTS v2 delivers frontier TTS at just $0.70 per generated hour. Available globally today as tts-rt-v2. Try it out: Turn the sound on and hear what’s possible. 🔊

Soniox

670,209 views • 2 months ago

This might be the beginning of a new field.. Researchers just reconstructed sounds from human brain activity using an fMRI and a generative AI model.

This might be the beginning of a new field.. Researchers just reconstructed sounds from human brain activity using an fMRI and a generative AI model.

Lior Alexander

208,406 views • 3 years ago

🏗️ Policy Adaptation from Foundation Model Feedback #CVPR2023 Instead of using foundation model as a pre-trained encoder (generator), we use it as a Teacher (discriminator) to tell where our policy did wrong and helps it adapts to new envs and tasks.

🏗️ Policy Adaptation from Foundation Model Feedback #CVPR2023 Instead of using foundation model as a pre-trained encoder (generator), we use it as a Teacher (discriminator) to tell where our policy did wrong and helps it adapts to new envs and tasks.

Xiaolong Wang

24,413 views • 3 years ago

The Segment Anything Model (SAM) by Meta AI is a step toward the first foundation model for image segmentation. SAM is capable of one-click segmentation of any object from photos or videos + zero-shot transfer to other segmentation tasks. Try the demo ➡️

The Segment Anything Model (SAM) by Meta AI is a step toward the first foundation model for image segmentation. SAM is capable of one-click segmentation of any object from photos or videos + zero-shot transfer to other segmentation tasks. Try the demo ➡️

AI at Meta

186,348 views • 3 years ago

Using monkey brain data to operate robotic arms! 🙉 Project Jenkins explores brain-computer interfaces by decoding neural activity into robotic movement and generating synthetic brain data from motion patterns. Using neural recordings from a macaque monkey named Jenkins, researchers developed models to decode brain signals into robotic arm movements and encode movement data into simulated neural activity. The project utilizes robotic arms and an interactive web console for real-time brain data generation from joystick inputs. The research paves the way for adaptive brain-computer interfaces by modeling brain activity more flexibly. Congrats to authors Dima Yanovsky, and Andrii Zahorodnii! 🦾

Using monkey brain data to operate robotic arms! 🙉 Project Jenkins explores brain-computer interfaces by decoding neural activity into robotic movement and generating synthetic brain data from motion patterns. Using neural recordings from a macaque monkey named Jenkins, researchers developed models to decode brain signals into robotic arm movements and encode movement data into simulated neural activity. The project utilizes robotic arms and an interactive web console for real-time brain data generation from joystick inputs. The research paves the way for adaptive brain-computer interfaces by modeling brain activity more flexibly. Congrats to authors Dima Yanovsky, and Andrii Zahorodnii! 🦾

Lukas Ziegler

21,637 views • 1 year ago

Neuralink just released a major new update on how it is training brain-computer interfaces Participants have now generated more than 50,000 HOURS of unlabeled neural data Neuralink is now pretraining neural encoders on thousands of hours of each participant’s brain activity before using them for cursor control And the results are already impressive: • Calibration time dropped from ~55 minutes per week to ~10 minutes for some participants • Multiple participants set new personal BCI speed records • Three participants surpassed the previous 10.39 bits/sec benchmark • One participant reached 11.32 bits/sec — a new BCI record • Some decoders stayed strong for weeks • One decoder was still usable more than 1.5 years later The most important part is what comes next Neuralink is moving toward foundation models for BCI...systems that can learn from massive amounts of neural data and eventually generalize across people and tasks The goal is to make brain-computer interfaces faster, more stable, require less calibration and eventually work more like a general-purpose platform Neuralink is going from controlling a cursor To building an interface between the brain and computers at scale

Neuralink just released a major new update on how it is training brain-computer interfaces Participants have now generated more than 50,000 HOURS of unlabeled neural data Neuralink is now pretraining neural encoders on thousands of hours of each participant’s brain activity before using them for cursor control And the results are already impressive: • Calibration time dropped from ~55 minutes per week to ~10 minutes for some participants • Multiple participants set new personal BCI speed records • Three participants surpassed the previous 10.39 bits/sec benchmark • One participant reached 11.32 bits/sec — a new BCI record • Some decoders stayed strong for weeks • One decoder was still usable more than 1.5 years later The most important part is what comes next Neuralink is moving toward foundation models for BCI...systems that can learn from massive amounts of neural data and eventually generalize across people and tasks The goal is to make brain-computer interfaces faster, more stable, require less calibration and eventually work more like a general-purpose platform Neuralink is going from controlling a cursor To building an interface between the brain and computers at scale

X Freeze

18,574 views • 25 days ago

Here is the first look Kendo AI V2! V2 adds the ability for teams to onboard and manage users, the ability to create fully customizable agents for any offer, sales cycle, or niche, and even adds a call review section for training and feedback + massive updates to our model

Here is the first look Kendo AI V2! V2 adds the ability for teams to onboard and manage users, the ability to create fully customizable agents for any offer, sales cycle, or niche, and even adds a call review section for training and feedback + massive updates to our model

Luke Alexander

13,339 views • 1 year ago

Gold on digital rails can open more than access. It can create new ways for people to enter, exit, learn, and participate in the world of finance. Most RWA tokens focus on providing exposure to an asset, but Ubuntu Tribe is building a community that empowers financial literacy and rewards participants. Together we rise!

Gold on digital rails can open more than access. It can create new ways for people to enter, exit, learn, and participate in the world of finance. Most RWA tokens focus on providing exposure to an asset, but Ubuntu Tribe is building a community that empowers financial literacy and rewards participants. Together we rise!

UTribe

10,343 views • 5 months ago

NEW: Higgs Audio V2 from BosonAI open, unified TTS model w/ voice cloning, beats GPT 4o mini tts and ElevenLabs v2 🔥 > Trained on 10M hours (speech, music, events) > Built on top of Llama 3.2 3B > Works real-time and on edge > Beats GPT-4o-mini-tts, ElevenLabs v2 in prosody & emotion Multi-speaker dialog > Zero-shot voice cloning 🤩 > Available on Hugging Face Kudos to folks at Boson AI for releasing such a brilliant work and all the details around the model! 🤗

NEW: Higgs Audio V2 from BosonAI open, unified TTS model w/ voice cloning, beats GPT 4o mini tts and ElevenLabs v2 🔥 > Trained on 10M hours (speech, music, events) > Built on top of Llama 3.2 3B > Works real-time and on edge > Beats GPT-4o-mini-tts, ElevenLabs v2 in prosody & emotion Multi-speaker dialog > Zero-shot voice cloning 🤩 > Available on Hugging Face Kudos to folks at Boson AI for releasing such a brilliant work and all the details around the model! 🤗

Vaibhav (VB) Srivastav

79,740 views • 1 year ago

Introducing Meta Perception Encoder: a vision encoder setting new standards in image & video tasks. It excels in zero-shot classification & retrieval, surpassing existing models. Learn more about Meta Perception Encoder, read the research paper, and download the code and dataset

Introducing Meta Perception Encoder: a vision encoder setting new standards in image & video tasks. It excels in zero-shot classification & retrieval, surpassing existing models. Learn more about Meta Perception Encoder, read the research paper, and download the code and dataset

AI at Meta

74,689 views • 1 year ago

We're building a surgical robot capable of reaching any brain region. The goal: a generalized neural interface to help solve any condition that originates in the brain.

We're building a surgical robot capable of reaching any brain region. The goal: a generalized neural interface to help solve any condition that originates in the brain.

Neuralink

22,119,998 views • 5 months ago

Nobel committee: stop overhyping the "AI works like the brain" story. At the announcement today of the 2026 Nobel Prize in Medicine, awarded to 3 scientists working on the mechanisms the brain uses to switch on, or off, the activity of individual nerve cells. A reporter asked whether the prize-winning brain research could help explain what happens inside AI neural networks. The panel answered: "Probably the similarities between the neural networks used in AI and the human brain shouldn't be exaggerated." ---- Full video on "Associated Press" YouTube channel, (link in comment)

Nobel committee: stop overhyping the "AI works like the brain" story. At the announcement today of the 2026 Nobel Prize in Medicine, awarded to 3 scientists working on the mechanisms the brain uses to switch on, or off, the activity of individual nerve cells. A reporter asked whether the prize-winning brain research could help explain what happens inside AI neural networks. The panel answered: "Probably the similarities between the neural networks used in AI and the human brain shouldn't be exaggerated." ---- Full video on "Associated Press" YouTube channel, (link in comment)

Rohan Paul

25,259 views • 23 days ago

Introducing Empathic Voice Interface 2 (EVI 2), our new voice-to-voice foundation model. EVI 2 merges language and voice into a single model trained specifically for emotional intelligence. You can try it and start building today.

Introducing Empathic Voice Interface 2 (EVI 2), our new voice-to-voice foundation model. EVI 2 merges language and voice into a single model trained specifically for emotional intelligence. You can try it and start building today.

Hume AI

165,645 views • 2 years ago

Introducing Natura-1, nature will's first general foundation model for the physical world. It is a single model that follows plain-language instructions zero-shot and carries them out with dexterous hands. - 50B MoE, pretrained on 1M hours of egocentric human experience. - Drives 20-DoF dexterous hands from plain language, zero-shot, in a 100ms loop. - Scaling human data from ~44K to 1M hours cut action MSE on held-out robot data by 67.7%. - Built on in-house data and compute infra, it handles fine-grained long-horizon tasks and a ~30-minute zero-shot interaction in one continuous take. The general model exists, and it can only get stronger.

Introducing Natura-1, nature will's first general foundation model for the physical world. It is a single model that follows plain-language instructions zero-shot and carries them out with dexterous hands. - 50B MoE, pretrained on 1M hours of egocentric human experience. - Drives 20-DoF dexterous hands from plain language, zero-shot, in a 100ms loop. - Scaling human data from ~44K to 1M hours cut action MSE on held-out robot data by 67.7%. - Built on in-house data and compute infra, it handles fine-grained long-horizon tasks and a ~30-minute zero-shot interaction in one continuous take. The general model exists, and it can only get stronger.

Ning Ding

21,884 views • 15 days ago

Introducing: Loyalty League v2 ⚡ Building on v1's key insights, we're thrilled to unveil a new standard, the Layer 2 of LST incentives 🧵

Introducing: Loyalty League v2 ⚡ Building on v1's key insights, we're thrilled to unveil a new standard, the Layer 2 of LST incentives 🧵

Hubra | Solana staking

22,967 views • 2 years ago

We are back. After one year of quiet building. Introducing GENE-26.5, our first robotic brain that takes a major step toward human-level capability. For years, robotics has struggled to learn from the world’s largest and valuable data source: Humans. Solving it means rethinking the whole stack from the ground up: - A robotics-native foundation model. - A 1:1 human-like robotic hand. - A noninvasive data collection glove for motion, force, and touch. - A simulator that turns weeks of experiments into minutes. GENE-26.5 is trained across language, vision, proprioception, tactile, and action. We designed a set of tasks to test how far we can go with this new paradigm. Fully autonomous, 1x speed, one model, same weights. (Enjoy with sound on) We are approaching the endgame for robotics. And this is just a beginning.

We are back. After one year of quiet building. Introducing GENE-26.5, our first robotic brain that takes a major step toward human-level capability. For years, robotics has struggled to learn from the world’s largest and valuable data source: Humans. Solving it means rethinking the whole stack from the ground up: - A robotics-native foundation model. - A 1:1 human-like robotic hand. - A noninvasive data collection glove for motion, force, and touch. - A simulator that turns weeks of experiments into minutes. GENE-26.5 is trained across language, vision, proprioception, tactile, and action. We designed a set of tasks to test how far we can go with this new paradigm. Fully autonomous, 1x speed, one model, same weights. (Enjoy with sound on) We are approaching the endgame for robotics. And this is just a beginning.

Genesis AI

2,749,405 views • 5 months ago

🔉 Introducing SAM Audio, the first unified model that isolates any sound from complex audio mixtures using text, visual, or span prompts. We’re sharing SAM Audio with the community, along with a perception encoder model, benchmarks and research papers, to empower others to explore new forms of expression and build applications that were previously out of reach. 🔗 Learn more:

🔉 Introducing SAM Audio, the first unified model that isolates any sound from complex audio mixtures using text, visual, or span prompts. We’re sharing SAM Audio with the community, along with a perception encoder model, benchmarks and research papers, to empower others to explore new forms of expression and build applications that were previously out of reach. 🔗 Learn more:

AI at Meta

1,253,770 views • 10 months ago