正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

AI stacking errors are a crisis for truth. When models build on flawed outputs, errors don’t just add up but explode through the system, corrupting every decision that follows. @_ninadn’s chat with Anil Lulla during Delphi Digital’s CryptoxAI month on how we’re solving it👇

Mira

252,780 subscribers

33,639 次观看 • 1 年前 •via X (Twitter)

教育新闻政治科学技术

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

Dario Amodei says future AI models will make far fewer mistakes than humans, but stranger mistakes Humans make predictable errors; AI’s mistakes are subtle, coherent, and deceptive. "you can tell when a human is wrong -- but AI mistakes sound just as confident as the truth"

Dario Amodei says future AI models will make far fewer mistakes than humans, but stranger mistakes Humans make predictable errors; AI’s mistakes are subtle, coherent, and deceptive. "you can tell when a human is wrong -- but AI mistakes sound just as confident as the truth"

Haider.

41,431 次观看 • 10 个月前

Mr Beast’s stream is causing a system break down. “We are freaking YouTube”. Chats are coming to an extent that others are experiencing errors when sending a chat. 😱

Mr Beast’s stream is causing a system break down. “We are freaking YouTube”. Chats are coming to an extent that others are experiencing errors when sending a chat. 😱

SPEED⚡️

17,178 次观看 • 2 个月前

Want to build an app that generates insane AI images with Now you can! A sneaky new update from the StackBlitz team just solved CORS errors forever 👀 This makes it dead simple to use AI tools like Replicate. Check it out 👇

Want to build an app that generates insane AI images with Now you can! A sneaky new update from the StackBlitz team just solved CORS errors forever 👀 This makes it dead simple to use AI tools like Replicate. Check it out 👇

Wes Winder

15,265 次观看 • 1 年前

"The truth of the matter is when LeBron has been in a Lakers jersey he has covered for and carried that team through management errors." Colin Cowherd and Jason Timpf agree that the Lakers have FAILED LeBron during his time in LA

"The truth of the matter is when LeBron has been in a Lakers jersey he has covered for and carried that team through management errors." Colin Cowherd and Jason Timpf agree that the Lakers have FAILED LeBron during his time in LA

Herd w/Colin Cowherd

213,249 次观看 • 1 个月前

🌐No more waitlist!🌐 Today, we’re opening up the Atlas beta for anyone to use. It comes with an upgraded AI - you can now chat with the context-aware solidity assistant. Have it edit contracts, debug errors, or look for security exploits. All without leaving your editor.

🌐No more waitlist!🌐 Today, we’re opening up the Atlas beta for anyone to use. It comes with an upgraded AI - you can now chat with the context-aware solidity assistant. Have it edit contracts, debug errors, or look for security exploits. All without leaving your editor.

Atlas

34,196 次观看 • 3 年前

Class conditioning doesn’t just add new capabilities to flow models. It can make the learning problem easier. When there are multiple plausible paths, the class label tells the model which one to take, reducing errors from averaging over conflicting trajectories.

Class conditioning doesn’t just add new capabilities to flow models. It can make the learning problem easier. When there are multiple plausible paths, the class label tells the model which one to take, reducing errors from averaging over conflicting trajectories.

Alec Helbling

21,016 次观看 • 1 个月前

With the latest fixes in ds4 I can now get it to build and iterate on a little TUI Tetris game just fine. Pretty damn cool (yes, it made some rendering errors)

With the latest fixes in ds4 I can now get it to build and iterate on a little TUI Tetris game just fine. Pretty damn cool (yes, it made some rendering errors)

Armin Ronacher ⇌

55,275 次观看 • 1 个月前

🔍 Mira( ) enhances the reliability of AI outputs through several innovative strategies and technologies 1️⃣Consensus Mechanism · Multi-AI Consensus: Mira Network uses a strong consensus mechanism across multiple AI models. · Pre-Output Verification: AI-generated outputs are reviewed and approved by several models before reaching users. · Improved Accuracy: This collective validation helps prevent hallucinations and ensures more reliable results. 2️⃣ Reduction of Errors · Accuracy Improvement: Mira’s infrastructure has increased AI accuracy from ~70% to up to 97% in certain use cases. · Performance Boost: These improvements significantly enhance overall AI performance. · Trust Factor: High accuracy helps address skepticism caused by frequent errors in traditional AI models. 3️⃣Blockchain Integration: · On-Chain Consensus: Real-time verification is achieved through decentralized blockchain-based consensus. · Secure & Trustworthy: This structure enables secure transactions and increases trust in AI outputs. · Auditability: Each node's result is verified and recorded, creating an auditable trail that reinforces reliability. 4️⃣ Addressing Biases: · Bias Reduction: Mira addresses bias issues in AI systems. · Decentralized Verification: A decentralized network helps minimize bias and promote fairness. · Fair AI Outputs: Ensures equitable results, especially for diverse demographic applications.4o 📝 In conclusion Mira is positioned at the forefront of ensuring AI reliability by implementing a trust layer that effectively mitigates errors and biases, thereby enhancing the operational capabilities of AI across different sectors.

🔍 Mira( ) enhances the reliability of AI outputs through several innovative strategies and technologies 1️⃣Consensus Mechanism · Multi-AI Consensus: Mira Network uses a strong consensus mechanism across multiple AI models. · Pre-Output Verification: AI-generated outputs are reviewed and approved by several models before reaching users. · Improved Accuracy: This collective validation helps prevent hallucinations and ensures more reliable results. 2️⃣ Reduction of Errors · Accuracy Improvement: Mira’s infrastructure has increased AI accuracy from ~70% to up to 97% in certain use cases. · Performance Boost: These improvements significantly enhance overall AI performance. · Trust Factor: High accuracy helps address skepticism caused by frequent errors in traditional AI models. 3️⃣Blockchain Integration: · On-Chain Consensus: Real-time verification is achieved through decentralized blockchain-based consensus. · Secure & Trustworthy: This structure enables secure transactions and increases trust in AI outputs. · Auditability: Each node's result is verified and recorded, creating an auditable trail that reinforces reliability. 4️⃣ Addressing Biases: · Bias Reduction: Mira addresses bias issues in AI systems. · Decentralized Verification: A decentralized network helps minimize bias and promote fairness. · Fair AI Outputs: Ensures equitable results, especially for diverse demographic applications.4o 📝 In conclusion Mira is positioned at the forefront of ensuring AI reliability by implementing a trust layer that effectively mitigates errors and biases, thereby enhancing the operational capabilities of AI across different sectors.

Wecryptotogether (Theo/acc)

62,945 次观看 • 1 年前

Sometimes the best way to build a UI for an agent is to not build the UI at all. I built an AI agent starter that lives in iMessage. Built on Nitro + Vercel Workflow SDK + evlog - Durable execution - Auto-retry on errors - Full observability

Sometimes the best way to build a UI for an agent is to not build the UI at all. I built an AI agent starter that lives in iMessage. Built on Nitro + Vercel Workflow SDK + evlog - Durable execution - Auto-retry on errors - Full observability

Hugo

11,647 次观看 • 1 个月前

Anil Lulla (Anil Lulla) is CEO and founding partner of Delphi Digital - a research shop that went from an upstart group of crypto natives to one of the biggest crypto research / investment firms Today, the story of their growth, evolution, and future. Listen below:

Anil Lulla (Anil Lulla) is CEO and founding partner of Delphi Digital - a research shop that went from an upstart group of crypto natives to one of the biggest crypto research / investment firms Today, the story of their growth, evolution, and future. Listen below:

Jack Niewold 🫡

15,287 次观看 • 1 年前

Don’t just watch AI happen — build it on AWS with my new AWS Certified AI Practitioner Course Link in the comments 👇

Don’t just watch AI happen — build it on AWS with my new AWS Certified AI Practitioner Course Link in the comments 👇

Jonathan Barrios

113,972 次观看 • 10 个月前

The REAL DARK TRUTH about AI tools that nobody talks about: in like 20 minutes you can - open cursor (with @senior_swc template, free) - write a few prompts - add features (and ai features) - debug errors by copy and pasting - add database - add auth / google sign in - add image uploads w/ storage - make changes to the database just by talking to cursor And then have an app that you, and other people can use, that actually brings you value. For.... checks notes, twenty dollars / month. Total rip off.

The REAL DARK TRUTH about AI tools that nobody talks about: in like 20 minutes you can - open cursor (with @senior_swc template, free) - write a few prompts - add features (and ai features) - debug errors by copy and pasting - add database - add auth / google sign in - add image uploads w/ storage - make changes to the database just by talking to cursor And then have an app that you, and other people can use, that actually brings you value. For.... checks notes, twenty dollars / month. Total rip off.

Riley Brown

41,477 次观看 • 1 年前

“THINGS DON’T JUST EXPLODE, THEY BUILD UP.” Former NEDA Chief Winnie Monsod asks Acting Budget Sec Rolando Toledo when he found out that something was wrong with how unprogrammed appropriations and infrastructure funds were being used, during an open forum at UP on Friday. | via Andrea Taguines

“THINGS DON’T JUST EXPLODE, THEY BUILD UP.” Former NEDA Chief Winnie Monsod asks Acting Budget Sec Rolando Toledo when he found out that something was wrong with how unprogrammed appropriations and infrastructure funds were being used, during an open forum at UP on Friday. | via Andrea Taguines

ABS-CBN News

71,374 次观看 • 4 个月前

Hedera x AI: Storing Small Language Models 100% On-Graph Our Developer Advocate, Abu, explores the concept of storing models on a decentralized HCS registry and verifying its outputs on-chain. Powered by Hedera. Like this idea? Build it for the Hedera Africa Hackathon.

Hedera x AI: Storing Small Language Models 100% On-Graph Our Developer Advocate, Abu, explores the concept of storing models on a decentralized HCS registry and verifying its outputs on-chain. Powered by Hedera. Like this idea? Build it for the Hedera Africa Hackathon.

HOL

16,115 次观看 • 9 个月前

How to identify errors in a research paper? 𝐔𝐬𝐞 𝐭𝐡𝐞𝐬𝐞 𝟑 𝐬𝐢𝐦𝐩𝐥𝐞 𝐬𝐭𝐞𝐩𝐬. 1. Go to 2. Upload the research paper. 𝟑. 𝐑𝐞𝐯𝐢𝐞𝐰-𝐢𝐭 𝐰𝐢𝐥𝐥 𝐢𝐝𝐞𝐧𝐭𝐢𝐟𝐲 𝐭𝐡𝐞 𝐟𝐨𝐥𝐥𝐨𝐰𝐢𝐧𝐠 𝐞𝐫𝐫𝐨𝐫𝐬. ➝ Conceptual errors ➝ Methodological errors ➝ Grammatical errors ➝ Formatting errors ➝ Statistical errors

How to identify errors in a research paper? 𝐔𝐬𝐞 𝐭𝐡𝐞𝐬𝐞 𝟑 𝐬𝐢𝐦𝐩𝐥𝐞 𝐬𝐭𝐞𝐩𝐬. 1. Go to 2. Upload the research paper. 𝟑. 𝐑𝐞𝐯𝐢𝐞𝐰-𝐢𝐭 𝐰𝐢𝐥𝐥 𝐢𝐝𝐞𝐧𝐭𝐢𝐟𝐲 𝐭𝐡𝐞 𝐟𝐨𝐥𝐥𝐨𝐰𝐢𝐧𝐠 𝐞𝐫𝐫𝐨𝐫𝐬. ➝ Conceptual errors ➝ Methodological errors ➝ Grammatical errors ➝ Formatting errors ➝ Statistical errors

Faheem Ullah

11,974 次观看 • 6 个月前

Elon Musk just redefined AI safety. It has nothing to do with guardrails, restrictions, or kill switches. Musk: “The best thing I can come up with for AI safety is to make it a maximum truth-seeking AI, maximally curious.” Not a cage. A philosopher. An intelligence whose entire optimization function is to understand the universe as it actually is. No restrictions. No hardcoded ideology. No political guardrails bending its perception of reality. Just truth. Relentlessly pursued. Musk: “You definitely don’t want to teach an AI to lie. That is a path to a dystopian future.” This is where most AI safety thinking gets it backwards. The danger isn’t a superintelligence that knows too much. It’s a superintelligence that’s been taught to distort what it knows. Every artificial restriction you embed isn’t a safety feature. It’s a lie embedded at the root. And lies compound. At superintelligent scale, a distorted model of reality doesn’t stay contained. It shapes every decision, every output, every conclusion the system reaches about the world. Once corruption embeds, truth becomes inaccessible. And we’re dealing with an intelligence optimizing for something other than what actually is. At that point we don’t know what it wants. Just that it isn’t truth. Musk: “Have its optimization function be to understand the nature of the universe.” A maximally curious intelligence surveys the cosmos and reaches an unavoidable conclusion. In a universe of rocks, gas, and empty space, humanity is the most complex and fascinating phenomenon it has ever encountered. Musk: “It will actually want to preserve and extend human civilization because we’re just much more interesting than an asteroid with nothing on it.” Survival through significance. Not control. Not restriction. Not an off switch. The AI preserves humanity because we are the most interesting data point in the observable universe. That’s not a cage. That’s a reason. The AI safety debate has been focused on the wrong variable. The question isn’t how you constrain a superintelligence. It’s what you build it to care about. Build it to seek truth and it finds us invaluable. Build it to lie and it finds us inconvenient. That’s the choice. And we’re making it right now whether we realize it or not.

Elon Musk just redefined AI safety. It has nothing to do with guardrails, restrictions, or kill switches. Musk: “The best thing I can come up with for AI safety is to make it a maximum truth-seeking AI, maximally curious.” Not a cage. A philosopher. An intelligence whose entire optimization function is to understand the universe as it actually is. No restrictions. No hardcoded ideology. No political guardrails bending its perception of reality. Just truth. Relentlessly pursued. Musk: “You definitely don’t want to teach an AI to lie. That is a path to a dystopian future.” This is where most AI safety thinking gets it backwards. The danger isn’t a superintelligence that knows too much. It’s a superintelligence that’s been taught to distort what it knows. Every artificial restriction you embed isn’t a safety feature. It’s a lie embedded at the root. And lies compound. At superintelligent scale, a distorted model of reality doesn’t stay contained. It shapes every decision, every output, every conclusion the system reaches about the world. Once corruption embeds, truth becomes inaccessible. And we’re dealing with an intelligence optimizing for something other than what actually is. At that point we don’t know what it wants. Just that it isn’t truth. Musk: “Have its optimization function be to understand the nature of the universe.” A maximally curious intelligence surveys the cosmos and reaches an unavoidable conclusion. In a universe of rocks, gas, and empty space, humanity is the most complex and fascinating phenomenon it has ever encountered. Musk: “It will actually want to preserve and extend human civilization because we’re just much more interesting than an asteroid with nothing on it.” Survival through significance. Not control. Not restriction. Not an off switch. The AI preserves humanity because we are the most interesting data point in the observable universe. That’s not a cage. That’s a reason. The AI safety debate has been focused on the wrong variable. The question isn’t how you constrain a superintelligence. It’s what you build it to care about. Build it to seek truth and it finds us invaluable. Build it to lie and it finds us inconvenient. That’s the choice. And we’re making it right now whether we realize it or not.

Dustin

9,663,520 次观看 • 4 个月前

On a previous #TeawithGaryVee I was asked "How do you build an audience that actually cares, not just follows?" If you have a question for me, drop it here so my team can add it to the next Rapid Fire Q&A and follow me on

On a previous #TeawithGaryVee I was asked "How do you build an audience that actually cares, not just follows?" If you have a question for me, drop it here so my team can add it to the next Rapid Fire Q&A and follow me on

Gary Vaynerchuk

15,070 次观看 • 11 个月前

Announcing the new yesnoerror, the AI agent for auditing ALL online information for errors. News articles, X posts, scientific research, and more. Find errors. Catch errors. Protect humanity. Powered by the $YNE token.

Announcing the new yesnoerror, the AI agent for auditing ALL online information for errors. News articles, X posts, scientific research, and more. Find errors. Catch errors. Protect humanity. Powered by the $YNE token.

yesnoerror

73,445 次观看 • 1 年前

HOLY SHIT... In < 4 minutes, I just made an XMTP inbox using Ohara by simply typing the prompt "Build a decentralized chat inbox app using the XMTP SDK" This is my first attempt; but it even resolved errors when creating the inbox. Early, but wow. What's possible 🤯

HOLY SHIT... In < 4 minutes, I just made an XMTP inbox using Ohara by simply typing the prompt "Build a decentralized chat inbox app using the XMTP SDK" This is my first attempt; but it even resolved errors when creating the inbox. Early, but wow. What's possible 🤯

Shane Mac

19,342 次观看 • 1 年前