Загрузка видео...

Не удалось загрузить видео

На главную

Right now, when you send a query to an LLM, it gets decrypted on the server. The LLM sees your data in plain text. Prof. Ajay Joshi (BU, CipherSonic AI ) on fully homomorphic encryption, which may be key for the future of AI privacy: how we can compute...

cerebras's profile picture

Cerebras

65,581 subscribers

293,561 просмотров • 2 месяцев назад •via X (Twitter)

ОбразованиеНовости и политикаНаука и технологии
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

Fully homomorphic encryption was invented in the 1980s. Why wasn't it adopted sooner? A 100,000x slowdown, driven by memory boundedness. Ajay Joshi from CipherSonic AI explains how his team got it down to less than 2x. (if this pattern sounds familar... LLM inference is memory-bound too. It's why wafer-scale exists.)
2:12

Fully homomorphic encryption was invented in the 1980s. Why wasn't it adopted sooner? A 100,000x slowdown, driven by memory boundedness. Ajay Joshi from CipherSonic AI explains how his team got it down to less than 2x. (if this pattern sounds familar... LLM inference is memory-bound too. It's why wafer-scale exists.)

Cerebras

57,733 просмотров • 3 месяцев назад

Encrypt is coming to Solana. Fully homomorphic encryption for the Solana Virtual Machine. Applications that compute on encrypted data without ever decrypting it. Encrypted Capital Markets.
0:32

Encrypt is coming to Solana. Fully homomorphic encryption for the Solana Virtual Machine. Applications that compute on encrypted data without ever decrypting it. Encrypted Capital Markets.

Encrypt

90,417 просмотров • 4 месяцев назад

Fully homomorphic encryption (FHE) lets you compute on encrypted data without ever decrypting it. It could change how we think about privacy, smart contracts, and AI. On Decypher, I spoke with Rand, founder of Now @zama, about where FHE stands today and what it makes possible. Listen below:
0:34

Fully homomorphic encryption (FHE) lets you compute on encrypted data without ever decrypting it. It could change how we think about privacy, smart contracts, and AI. On Decypher, I spoke with Rand, founder of Now @zama, about where FHE stands today and what it makes possible. Listen below:

Jack Niewold 🫡

35,938 просмотров • 1 год назад

Yes, your data is encrypted in transit and at rest. But nobody talks about this. It still gets decrypted server-side, actually, to process it, even with all the safeguards. FHE lets you compute on encrypted data without ever decrypting it. Wild. Is it 1000x slower rn? Yeah, Will it work for everything? Maybe not, but for medical records, financial data, proprietary code? This is the future. Explained! Enjoy.
1:41

Yes, your data is encrypted in transit and at rest. But nobody talks about this. It still gets decrypted server-side, actually, to process it, even with all the safeguards. FHE lets you compute on encrypted data without ever decrypting it. Wild. Is it 1000x slower rn? Yeah, Will it work for everything? Maybe not, but for medical records, financial data, proprietary code? This is the future. Explained! Enjoy.

I D R I S

16,630 просмотров • 8 месяцев назад

we sent 2 PMs to Microsoft Build with one mission: whoever gets the most photos with the Wafer-Scale Engine in 10 minuntes, wins. for anyone who hasn't met it in person — the WSE is the largest chip ever built, and it powers the fastest AI. It tends to draw a crowd!
1:01

we sent 2 PMs to Microsoft Build with one mission: whoever gets the most photos with the Wafer-Scale Engine in 10 minuntes, wins. for anyone who hasn't met it in person — the WSE is the largest chip ever built, and it powers the fastest AI. It tends to draw a crowd!

Cerebras

13,039 просмотров • 2 месяцев назад

Perplexity CEO Aravind Srinivas on the biggest threat to the data center industry: It's not competition. It's not regulation. It's decentralisation. "The biggest threat to a data center is if the intelligence can be packed locally on a chip that's running on the device and then there's no need to inference all of it on like one centralized data center." He outlines how this could work in practice. Personalisation doesn't necessarily require on-device model training. Retrieval augmented generation, tool calls, and local data can already tailor AI to individual users. But the real unlock? Test time training. Aravind Srinivas describes a future where AI lives on your device, watches how you work and gradually automates your repetitive tasks. "Imagine we crack test time training where the AI watches tasks you repeatedly do on your local system, adapts to you over time and starts automating a lot of the things you do." The key insight: in this model, the intelligence belongs to you. It's your data, your device, your personalised AI brain. And if that future arrives, the economics of centralised infrastructure start to collapse. "That really disrupts the whole data center industry. It doesn't make sense to spend all this money, 500 billion, 5 trillion, whatever on building all the centralized data centers across the world that do a lot of the intelligence workloads for people." The companies spending trillions on centralised infrastructure may want to rethink where intelligence actually needs to live.
1:15

Perplexity CEO Aravind Srinivas on the biggest threat to the data center industry: It's not competition. It's not regulation. It's decentralisation. "The biggest threat to a data center is if the intelligence can be packed locally on a chip that's running on the device and then there's no need to inference all of it on like one centralized data center." He outlines how this could work in practice. Personalisation doesn't necessarily require on-device model training. Retrieval augmented generation, tool calls, and local data can already tailor AI to individual users. But the real unlock? Test time training. Aravind Srinivas describes a future where AI lives on your device, watches how you work and gradually automates your repetitive tasks. "Imagine we crack test time training where the AI watches tasks you repeatedly do on your local system, adapts to you over time and starts automating a lot of the things you do." The key insight: in this model, the intelligence belongs to you. It's your data, your device, your personalised AI brain. And if that future arrives, the economics of centralised infrastructure start to collapse. "That really disrupts the whole data center industry. It doesn't make sense to spend all this money, 500 billion, 5 trillion, whatever on building all the centralized data centers across the world that do a lot of the intelligence workloads for people." The companies spending trillions on centralised infrastructure may want to rethink where intelligence actually needs to live.

Big Brain AI

90,276 просмотров • 6 месяцев назад

What if your personal data became an asset you actually owned? In this May 2026 AMA, Matrix AI Network explores one of the most important questions facing AI, Web3, and the future digital economy: Who owns the data that powers artificial intelligence? As AI models continue to improve, the next major bottleneck is no longer computing power—it is access to high-quality, verified human data. The discussion covers how personal biodata can become a valuable digital asset, how data quality is measured, and why future AI companies may depend directly on users as data providers. Topics include: ✅ Biodata as a new asset class ✅ AI-grade data quality scoring ✅ Wearable devices for raw data ✅ Data ownership and monetization models ✅ Preventing fake data and device spoofing ✅ Real-world applications in healthcare, sports science and mental wellness ✅ Hypnus and the future biodata ecosystem ✅ Project direction The future may not belong to companies that own data. It belongs to the people who create it. Watch now 👇 #AI #Web3 #DataOwnership
29:51

What if your personal data became an asset you actually owned? In this May 2026 AMA, Matrix AI Network explores one of the most important questions facing AI, Web3, and the future digital economy: Who owns the data that powers artificial intelligence? As AI models continue to improve, the next major bottleneck is no longer computing power—it is access to high-quality, verified human data. The discussion covers how personal biodata can become a valuable digital asset, how data quality is measured, and why future AI companies may depend directly on users as data providers. Topics include: ✅ Biodata as a new asset class ✅ AI-grade data quality scoring ✅ Wearable devices for raw data ✅ Data ownership and monetization models ✅ Preventing fake data and device spoofing ✅ Real-world applications in healthcare, sports science and mental wellness ✅ Hypnus and the future biodata ecosystem ✅ Project direction The future may not belong to companies that own data. It belongs to the people who create it. Watch now 👇 #AI #Web3 #DataOwnership

Matrix AI Network

60,855 просмотров • 2 месяцев назад

"Claude Code is the first AI tool that can actually recreate how a data scientist thinks." Here's my new episode with Sumeet Marwaha (Head of Data at Brex) on exactly how to build an AI analyst with Claude Code. We covered: ✅ 3 queries to build a data analyst MCP ✅ How to give the analyst both data + context ✅ Exclusive Brex data on which AI tools are crushing it in startups and enterprise Some quotes from Sumeet: "I've set up dashboards many times in my career. They end up getting ignored. Claude will always read it. Claude will always ask the right questions." "You have to set up Claude Code to not just look at the data. It should be able to search Slack, find the incident, and understand why your metrics look wrong." "This tool is absolutely crushing it based on our data. It's both the startup AND enterprise coding tool of choice." 📌 Watch now:
0:46

"Claude Code is the first AI tool that can actually recreate how a data scientist thinks." Here's my new episode with Sumeet Marwaha (Head of Data at Brex) on exactly how to build an AI analyst with Claude Code. We covered: ✅ 3 queries to build a data analyst MCP ✅ How to give the analyst both data + context ✅ Exclusive Brex data on which AI tools are crushing it in startups and enterprise Some quotes from Sumeet: "I've set up dashboards many times in my career. They end up getting ignored. Claude will always read it. Claude will always ask the right questions." "You have to set up Claude Code to not just look at the data. It should be able to search Slack, find the incident, and understand why your metrics look wrong." "This tool is absolutely crushing it based on our data. It's both the startup AND enterprise coding tool of choice." 📌 Watch now:

Peter Yang

39,600 просмотров • 7 месяцев назад

Today we're launching Amplitude MCP MCP was made for data analytics. Ask a vague question and have an AI model iteratively query Amplitude to find you insights. Amplitude MCP exposes all of Amplitude's functionality so an agent can interact with it directly. It allows you to use Amplitude without having to know anything about Amplitude's user interface or your data taxonomy. It uses the intelligence of off the shelf AI agents, including Claude, Gemini, and Cursor, to query Amplitude data you already have to find answers. It leverages all of your existing Amplitude data and how it's used to navigate through large complex behavioral datasets. You can have conversations about your product data directly inside any AI model with MCP. It can create Amplitude dashboards and notebooks with what it finds and even make recommendations. This is the best MCP use case I've ever seen. Watching an AI agent think, reason, query Amplitude, and then repeat that process iteratively is magical. See it in action:
3:08

Today we're launching Amplitude MCP MCP was made for data analytics. Ask a vague question and have an AI model iteratively query Amplitude to find you insights. Amplitude MCP exposes all of Amplitude's functionality so an agent can interact with it directly. It allows you to use Amplitude without having to know anything about Amplitude's user interface or your data taxonomy. It uses the intelligence of off the shelf AI agents, including Claude, Gemini, and Cursor, to query Amplitude data you already have to find answers. It leverages all of your existing Amplitude data and how it's used to navigate through large complex behavioral datasets. You can have conversations about your product data directly inside any AI model with MCP. It can create Amplitude dashboards and notebooks with what it finds and even make recommendations. This is the best MCP use case I've ever seen. Watching an AI agent think, reason, query Amplitude, and then repeat that process iteratively is magical. See it in action:

Spenser Skates

32,332 просмотров • 10 месяцев назад

Consistent Hashing Explained Simply Let's break down consistent hashing in a way that's easy to get. It's a handy technique for distributing data evenly across servers. And if you're more of a visual learner, don't miss the attached video which brings this concept to life. Here's how it works: We don't just hash our data keys; we also hash the server names using the same method. This puts everything on a scale, known as the hash space, which we think of as a big circular ring. So, imagine we've got a few servers. We hash them and place them on this ring. We do the same with our data, hashing it by key. To figure out where to store a piece of data, we travel clockwise on the ring from where the data lands until we bump into a server. Like, data key 0 might end up on server 0, and key 1 on server 1. Adding a new server? Only the data near the new server needs to shift. And if a server has to go? Only the data it was holding needs a new spot. But there's a little challenge: sometimes, data doesn't spread out evenly. The fix? Virtual nodes. They're like mini-replicas of a server placed around the ring. The more virtual nodes we have, the better the balance. But there's a catch: more virtual nodes mean more memory usage for tracking them, and it can make the whole system a bit more complex to manage. – Subscribe to our weekly newsletter to get a Free System Design PDF (158 pages):
2:53

Consistent Hashing Explained Simply Let's break down consistent hashing in a way that's easy to get. It's a handy technique for distributing data evenly across servers. And if you're more of a visual learner, don't miss the attached video which brings this concept to life. Here's how it works: We don't just hash our data keys; we also hash the server names using the same method. This puts everything on a scale, known as the hash space, which we think of as a big circular ring. So, imagine we've got a few servers. We hash them and place them on this ring. We do the same with our data, hashing it by key. To figure out where to store a piece of data, we travel clockwise on the ring from where the data lands until we bump into a server. Like, data key 0 might end up on server 0, and key 1 on server 1. Adding a new server? Only the data near the new server needs to shift. And if a server has to go? Only the data it was holding needs a new spot. But there's a little challenge: sometimes, data doesn't spread out evenly. The fix? Virtual nodes. They're like mini-replicas of a server placed around the ring. The more virtual nodes we have, the better the balance. But there's a catch: more virtual nodes mean more memory usage for tracking them, and it can make the whole system a bit more complex to manage. – Subscribe to our weekly newsletter to get a Free System Design PDF (158 pages):

Sahn Lam

36,787 просмотров • 2 лет назад

🚨GEOFFREY HINTON: WE DON'T KNOW EXACTLY HOW AI WORKS AI Pioneer Geoffrey Hinton: "We have a rough idea of what it's doing, but when it gets complicated, we don't know what's going on, similar to our understanding of the brain." 60 Minutes: "What do you mean we don't know exactly how it works? It was designed by people." Hinton: "We designed the learning algorithm, like the principle of evolution. But when it interacts with data, it creates neural networks we don't fully understand." Source: 60 Minutes, Newsy on IG
0:33

🚨GEOFFREY HINTON: WE DON'T KNOW EXACTLY HOW AI WORKS AI Pioneer Geoffrey Hinton: "We have a rough idea of what it's doing, but when it gets complicated, we don't know what's going on, similar to our understanding of the brain." 60 Minutes: "What do you mean we don't know exactly how it works? It was designed by people." Hinton: "We designed the learning algorithm, like the principle of evolution. But when it interacts with data, it creates neural networks we don't fully understand." Source: 60 Minutes, Newsy on IG

Mario Nawfal

502,988 просмотров • 2 лет назад

#WATCH | India AI Impact Summit 2026 | Delhi: Founder Chairman and CEO of Sampark Foundation & former CEO of HCL Technologies, Vineet Nayar says, "...From an employment point of view I think it is very important for us to understand that Indian companies, including Indian IT companies, are going to be profit-driven and therefore if you believe that they are going to create employment you must be dreaming. Therefore, the question is how do we create employment in this environment, and that employment comes from mass scale startups, which is what this government has already doing. So, how do we create new sets of people who are trying to solve new sets of problems not new sets of technology and if we do that we will get it right. I think we as Indians have to be very careful on who does data belong to and that is the debate we have a problem with. The LLM models which exist worldwide are far superior than the Indian models. Unfortunately, in India, we never develop products, so therefore we do not have SLMs and LLMs which are world-class. On one side, we have global LLM products which are coming to India and trading on our Indian data. Should we allowed that or should we not allowed that? But on the other side if we don't allow that then we have the data but we don't have the LLM models. So, how do we encourage technology completely to develop the LLM models. This needs radicals strategic thinking and a very important aspect otherwise we will either give up a data. So, I think it's a very critical aspect for us to think about - who does this data belong, what is the kind of incentives we are going to give to develop LLM technologies or SLM technologies fast so that we train on our data otherwise an LLM will come in with our data and we'll immediately see return and we'll celebrate and we will do all these kind of press releases but the India will lose a competitive advantage on something which is very critical for the next decade."
5:37

#WATCH | India AI Impact Summit 2026 | Delhi: Founder Chairman and CEO of Sampark Foundation & former CEO of HCL Technologies, Vineet Nayar says, "...From an employment point of view I think it is very important for us to understand that Indian companies, including Indian IT companies, are going to be profit-driven and therefore if you believe that they are going to create employment you must be dreaming. Therefore, the question is how do we create employment in this environment, and that employment comes from mass scale startups, which is what this government has already doing. So, how do we create new sets of people who are trying to solve new sets of problems not new sets of technology and if we do that we will get it right. I think we as Indians have to be very careful on who does data belong to and that is the debate we have a problem with. The LLM models which exist worldwide are far superior than the Indian models. Unfortunately, in India, we never develop products, so therefore we do not have SLMs and LLMs which are world-class. On one side, we have global LLM products which are coming to India and trading on our Indian data. Should we allowed that or should we not allowed that? But on the other side if we don't allow that then we have the data but we don't have the LLM models. So, how do we encourage technology completely to develop the LLM models. This needs radicals strategic thinking and a very important aspect otherwise we will either give up a data. So, I think it's a very critical aspect for us to think about - who does this data belong, what is the kind of incentives we are going to give to develop LLM technologies or SLM technologies fast so that we train on our data otherwise an LLM will come in with our data and we'll immediately see return and we'll celebrate and we will do all these kind of press releases but the India will lose a competitive advantage on something which is very critical for the next decade."

ANI

18,753 просмотров • 6 месяцев назад

our biggest focus for the next 2 months at kled is trust & safety. so we built a fully public audit trail for every single upload on the app, updates automatically the second it's made. the moment we send data to an AI lab, they can search the hash ID of any image, video, doc, or audio file and see exactly: - the tos + privacy policy the user signed - every compliance check we ran - anonymized kyc proof - confirmation the user got paid - a full timestamp of the entire interaction start to finish any ai lab can now verify the legitimacy of any dataset in seconds, making us the first data marketplace in history to publicize its data audit records. consumer identities are fully anonymized and hidden, no users will be exposed in the process.
0:49

our biggest focus for the next 2 months at kled is trust & safety. so we built a fully public audit trail for every single upload on the app, updates automatically the second it's made. the moment we send data to an AI lab, they can search the hash ID of any image, video, doc, or audio file and see exactly: - the tos + privacy policy the user signed - every compliance check we ran - anonymized kyc proof - confirmation the user got paid - a full timestamp of the entire interaction start to finish any ai lab can now verify the legitimacy of any dataset in seconds, making us the first data marketplace in history to publicize its data audit records. consumer identities are fully anonymized and hidden, no users will be exposed in the process.

Avi Patel

44,231 просмотров • 3 месяцев назад

Atomic Mail is the first email product to run AI features through Chrome's built-in Gemini Nano, with all processing happening directly on your device. > No data leaves your machine. No server calls. Zero cost. > Toggle "Local LLM" on in Privacy Center > Security Options. > It’s now available to all Chrome Desktop users.
0:30

Atomic Mail is the first email product to run AI features through Chrome's built-in Gemini Nano, with all processing happening directly on your device. > No data leaves your machine. No server calls. Zero cost. > Toggle "Local LLM" on in Privacy Center > Security Options. > It’s now available to all Chrome Desktop users.

🚨 AI News | TestingCatalog

22,720 просмотров • 4 месяцев назад

We deployed a fully private AI agent on NuNet in under 5 minutes 🚀 OpenClaw🦞 running Qwen through ollama , one of the hottest open source model families right now, entirely on decentralized compute. No cloud. No API keys. No data leaving the machine. This is what private AI looks like when you actually build it instead of just talking about it. Your model. Your hardware. Your rules. Full walkthrough showing exactly how it works: What should we deploy next?
12:59

We deployed a fully private AI agent on NuNet in under 5 minutes 🚀 OpenClaw🦞 running Qwen through ollama , one of the hottest open source model families right now, entirely on decentralized compute. No cloud. No API keys. No data leaving the machine. This is what private AI looks like when you actually build it instead of just talking about it. Your model. Your hardware. Your rules. Full walkthrough showing exactly how it works: What should we deploy next?

NuNet 🌐

87,604 просмотров • 5 месяцев назад

Alexandr Wang, founder of Scale AI, on why the most overlooked layer of AI became its biggest competitive advantage: Wang left MIT in 2016 to start Scale AI, becoming the world's youngest self-made billionaire by betting on AI's unglamorous infrastructure layer while others chased flashier opportunities. His framework for understanding AI is simple. There are three major components: "You have compute, which folks like Nvidia work on. You have algorithms which folks like OpenAI or Google or Anthropic build. And then you have data." Compute had Nvidia. Algorithms had the major labs. Data had no one. That's exactly where Wang and his team made their bet: "We recognized early on that data was going to be a limiter. And there was nobody really working on it." What looked like a gap in the market turned out to be the foundation of it. Years later, Wang says every serious player in AI has arrived at the same conclusion: "Fast forward today, everybody I talked to recognizes that data is gold. Whether it's our customers, the major labs like the OpenAIs or the Metas of the world, or large enterprises like Morgan Stanley or Mastercard, or our partnership between Cisco and Scale." Alexandr Wang continues: "Data is going to be what will differentiate and power the next generation of powerful AI capabilities."
0:46

Alexandr Wang, founder of Scale AI, on why the most overlooked layer of AI became its biggest competitive advantage: Wang left MIT in 2016 to start Scale AI, becoming the world's youngest self-made billionaire by betting on AI's unglamorous infrastructure layer while others chased flashier opportunities. His framework for understanding AI is simple. There are three major components: "You have compute, which folks like Nvidia work on. You have algorithms which folks like OpenAI or Google or Anthropic build. And then you have data." Compute had Nvidia. Algorithms had the major labs. Data had no one. That's exactly where Wang and his team made their bet: "We recognized early on that data was going to be a limiter. And there was nobody really working on it." What looked like a gap in the market turned out to be the foundation of it. Years later, Wang says every serious player in AI has arrived at the same conclusion: "Fast forward today, everybody I talked to recognizes that data is gold. Whether it's our customers, the major labs like the OpenAIs or the Metas of the world, or large enterprises like Morgan Stanley or Mastercard, or our partnership between Cisco and Scale." Alexandr Wang continues: "Data is going to be what will differentiate and power the next generation of powerful AI capabilities."

Big Brain AI

11,442 просмотров • 3 месяцев назад

🚨 FRANCOIS CHOLLET SAYS THE CURRENT LLM STACK IS THE WRONG PATH TO AGI Interviewer: “When will we accomplish the first definition of AGI?” Chollet confirms we are already on that exact trajectory, pointing out that: "Current technology can fully automate at human level or beyond any domain where you have verifiable rewards, right? And code... code being the first one." He then brings up the problem with LLMs, stating that: "It's possible in principle to build something that looks a lot like AGI on top of the LLM stack... I do believe, however, this would be the wrong thing to do because it would be very inefficient." Looking ahead to what the final architecture will actually require, he explains that: "AI research will have to trend towards not just efficiency, but in fact optimality over time. And for this reason, future AI in a few decades... it's not going to be this harness on top of a reasoning model on top of a base LLM."
1:41

🚨 FRANCOIS CHOLLET SAYS THE CURRENT LLM STACK IS THE WRONG PATH TO AGI Interviewer: “When will we accomplish the first definition of AGI?” Chollet confirms we are already on that exact trajectory, pointing out that: "Current technology can fully automate at human level or beyond any domain where you have verifiable rewards, right? And code... code being the first one." He then brings up the problem with LLMs, stating that: "It's possible in principle to build something that looks a lot like AGI on top of the LLM stack... I do believe, however, this would be the wrong thing to do because it would be very inefficient." Looking ahead to what the final architecture will actually require, he explains that: "AI research will have to trend towards not just efficiency, but in fact optimality over time. And for this reason, future AI in a few decades... it's not going to be this harness on top of a reasoning model on top of a base LLM."

Chris

15,578 просмотров • 4 месяцев назад