Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

Inkling Thinking Machines 1-bit quant by Unsloth AI running 30-40 TPS 🤯 video sped up, feel free to jump at the end to check outputs also llama.cpp webui is 😍 has reasoning slider, html preview, mcp and multimodal support

mervenoyann's profile picture

merve

89,565 subscribers

25,297 görüntüleme • 26 gün önce •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 Yorum

Yorum bulunmuyor

Orijinal gönderinin yorumları burada görünecek

Benzer Videolar

Fuck it, 685B parameter, DeepSeek V3 0324 running locally on M3 Ultra, fully private 🔥 Powered by llama.cpp & dynamic quants from Unsloth AI ⚡ Step 1: brew install llama.cpp Step 2: llama-cli -hf unsloth/DeepSeek-V3-0324-GGUF:Q2_K_XL That's it! 🤗 Honestly a bit surreal to be able to chat with such a chunky model at the touch of the keyboard - future is going to be wild!!
5:00

Fuck it, 685B parameter, DeepSeek V3 0324 running locally on M3 Ultra, fully private 🔥 Powered by llama.cpp & dynamic quants from Unsloth AI ⚡ Step 1: brew install llama.cpp Step 2: llama-cli -hf unsloth/DeepSeek-V3-0324-GGUF:Q2_K_XL That's it! 🤗 Honestly a bit surreal to be able to chat with such a chunky model at the touch of the keyboard - future is going to be wild!!

Vaibhav (VB) Srivastav

168,141 görüntüleme • 1 yıl önce

Connecting Claude with web search is magic! Finally got the chance to play with MCP servers. I hooked up Claude Desktop with GitHub & Brave web search MCP servers for staying up to date with AI news. I also asked Claude to create a new GitHub repo and log the news there. This exceeded my expectations. Check the full video (at 2x) for the first demo. Let me know if you are interested in a full tutorial on how to set up the MCP servers.
6:37

Connecting Claude with web search is magic! Finally got the chance to play with MCP servers. I hooked up Claude Desktop with GitHub & Brave web search MCP servers for staying up to date with AI news. I also asked Claude to create a new GitHub repo and log the news there. This exceeded my expectations. Check the full video (at 2x) for the first demo. Let me know if you are interested in a full tutorial on how to set up the MCP servers.

elvis

41,804 görüntüleme • 1 yıl önce

We've just unveiled ERNIE 4.5 & X1! 🚀 As a deep-thinking reasoning model with multimodal capabilities, ERNIE X1 delivers performance on par with DeepSeek R1 at only half the price. Meanwhile, ERNIE 4.5 is our latest foundation model and new-generation native multimodal model. Plus, our AI chatbot ERNIE Bot has now been made free to individual users ahead of schedule. Both models are now freely accessible to all ERNIE Bot users via its official website:
0:20

We've just unveiled ERNIE 4.5 & X1! 🚀 As a deep-thinking reasoning model with multimodal capabilities, ERNIE X1 delivers performance on par with DeepSeek R1 at only half the price. Meanwhile, ERNIE 4.5 is our latest foundation model and new-generation native multimodal model. Plus, our AI chatbot ERNIE Bot has now been made free to individual users ahead of schedule. Both models are now freely accessible to all ERNIE Bot users via its official website:

Baidu Inc.

1,712,153 görüntüleme • 1 yıl önce

100,000 users in < 5 months 🤯🤯🤯 next stop: 1 million users! Lightning AI Studio adoption by top AI labs and enterprises has been truly humbling... The 100,000th user gets free Pro tier for 3 months 👇🏼👇🏼 Congrats to Suhas Nandiraju at Crosby Health, the 100,000th user! who will get free Pro tier for 3 months! If you want to jump on the rocket ship and work on the next generation AI development platform, we're hiring in NYC, SF and London!
0:46

100,000 users in < 5 months 🤯🤯🤯 next stop: 1 million users! Lightning AI Studio adoption by top AI labs and enterprises has been truly humbling... The 100,000th user gets free Pro tier for 3 months 👇🏼👇🏼 Congrats to Suhas Nandiraju at Crosby Health, the 100,000th user! who will get free Pro tier for 3 months! If you want to jump on the rocket ship and work on the next generation AI development platform, we're hiring in NYC, SF and London!

William Falcon ⚡️

45,442 görüntüleme • 2 yıl önce

Dreamina Seedance 2.5 has officially dropped on Dreamina, the native AI creation hub by ByteDance. 2 massive upgrades that impressed me the most: • Throw in up to 50 multimodal reference assets (images, videos, and audio) in just one single generation. • Generate up to 30-second continuous video outputs for deeper, much more seamless storytelling. When you pair these capabilities with Dreamina's native editing workflow, multilingual support, and the broader Seedance ecosystem, we are looking at a serious leap toward actual production-grade AI video. For anyone looking to build on top of Seedance, Dreamina is my go-to recommendation right now. It houses the complete Seedance lineup, ranging from Seedance 2.0 Mini up to Seedance 2.5, offering creator-friendly pricing and standing out as one of the most cost-effective official routes to test, tweak, and scale your production. Less friction switching between tools. More focus on actual creation. #Dreamina #Seedance25 #AIVideo #ByteDance
0:30

Dreamina Seedance 2.5 has officially dropped on Dreamina, the native AI creation hub by ByteDance. 2 massive upgrades that impressed me the most: • Throw in up to 50 multimodal reference assets (images, videos, and audio) in just one single generation. • Generate up to 30-second continuous video outputs for deeper, much more seamless storytelling. When you pair these capabilities with Dreamina's native editing workflow, multilingual support, and the broader Seedance ecosystem, we are looking at a serious leap toward actual production-grade AI video. For anyone looking to build on top of Seedance, Dreamina is my go-to recommendation right now. It houses the complete Seedance lineup, ranging from Seedance 2.0 Mini up to Seedance 2.5, offering creator-friendly pricing and standing out as one of the most cost-effective official routes to test, tweak, and scale your production. Less friction switching between tools. More focus on actual creation. #Dreamina #Seedance25 #AIVideo #ByteDance

Jaynit Makwana

34,704 görüntüleme • 8 gün önce

🚨 Surfpool v0.7.0 is live, and the machines are taking control 🤖 Surfpool's MCP now has full access to the RPC layer — and we’re not entirely sure what it can’t do anymore 😂 This makes Surfpool the perfect companion for devs vibing with our friends at SendAI 😍 — big shoutout to noah gundotra for connecting the dots 💥 As usual, we’ve also added support for new RPC endpoints, bug fixes, and DX improvements. S/O arthur FX Michael Turbin3 for the contributions ❤️
1:46

🚨 Surfpool v0.7.0 is live, and the machines are taking control 🤖 Surfpool's MCP now has full access to the RPC layer — and we’re not entirely sure what it can’t do anymore 😂 This makes Surfpool the perfect companion for devs vibing with our friends at SendAI 😍 — big shoutout to noah gundotra for connecting the dots 💥 As usual, we’ve also added support for new RPC endpoints, bug fixes, and DX improvements. S/O arthur FX Michael Turbin3 for the contributions ❤️

ludo

12,411 görüntüleme • 1 yıl önce

Andrej Karpathy just explained the 5 shifts turning LLMs into agentic systems. 00:00 - Memory turns chat into personal AI 06:41 - Multimodal AI reads the world 16:58 - Thinking models solve harder tasks 24:51 - Search makes LLMs live 30:58 - Tools turn LLMs into workers Most people are still treating LLMs like chatbots. Karpathy is showing the full stack: Memory → Vision → Reasoning → Search → Tools Prompting is the old workflow. Agentic systems are the new one. This 40-minute talk is worth more than most paid AI agent courses. Bookmark and watch it before everyone catches up. Then read how to turn LLMs into self-improving agent loops below
42:43

Andrej Karpathy just explained the 5 shifts turning LLMs into agentic systems. 00:00 - Memory turns chat into personal AI 06:41 - Multimodal AI reads the world 16:58 - Thinking models solve harder tasks 24:51 - Search makes LLMs live 30:58 - Tools turn LLMs into workers Most people are still treating LLMs like chatbots. Karpathy is showing the full stack: Memory → Vision → Reasoning → Search → Tools Prompting is the old workflow. Agentic systems are the new one. This 40-minute talk is worth more than most paid AI agent courses. Bookmark and watch it before everyone catches up. Then read how to turn LLMs into self-improving agent loops below

Lunar

348,152 görüntüleme • 7 gün önce

🚀 Just launched: The GitHub Models repo integration is in public preview. ✨ Now you can: ➡️ Version, review, and collaborate on .prompt.yml files just like code. ➡️ Compare outputs side-by-side from 40+ leading LLMs. ➡️ Easily set up with a single API key for FREE. ➡️ Benefit from enterprise-ready governance and security on GitHub and Azure. Ready to start building efficiently with AI models? Get started. 👇
2:06

🚀 Just launched: The GitHub Models repo integration is in public preview. ✨ Now you can: ➡️ Version, review, and collaborate on .prompt.yml files just like code. ➡️ Compare outputs side-by-side from 40+ leading LLMs. ➡️ Easily set up with a single API key for FREE. ➡️ Benefit from enterprise-ready governance and security on GitHub and Azure. Ready to start building efficiently with AI models? Get started. 👇

GitHub

37,934 görüntüleme • 1 yıl önce

Check out the latest advert by the Scottish Electoral Commission. Its asking people to remember to register their vote for the upcoming Scottish elections. Let's have a look at the kind of people they are targeting 👇🏻 Lefty students (1 foreign), oldies (White, coz theyre old), 2 neighbours (1 foreign), a multicultural male netball team with 3 more foreigners... and another 2 more foreigners at the end. This woke bullshit has to end. Native White Scots are an afterthought to these crackpots running the country. The Scottish Electoral Commission is supposedly an "independent regulator" and not controlled directly by, (although it is accountable to) the Scottish government, and also gets its funding from them. Makes sense. Shut it all down.
0:30

Check out the latest advert by the Scottish Electoral Commission. Its asking people to remember to register their vote for the upcoming Scottish elections. Let's have a look at the kind of people they are targeting 👇🏻 Lefty students (1 foreign), oldies (White, coz theyre old), 2 neighbours (1 foreign), a multicultural male netball team with 3 more foreigners... and another 2 more foreigners at the end. This woke bullshit has to end. Native White Scots are an afterthought to these crackpots running the country. The Scottish Electoral Commission is supposedly an "independent regulator" and not controlled directly by, (although it is accountable to) the Scottish government, and also gets its funding from them. Makes sense. Shut it all down.

SnoopDeg

12,515 görüntüleme • 4 ay önce

Darkbloom just completed a major network upgrade! BIG UPDATE: We’re moving from Research Preview to Public Alpha. In the last month: - 1000s of provider signups, 250 live providers at peak - 600M+ tokens served With this upgrade, performance is up 30–200% across key metrics like TTFT, TPS, and total tokens served. The goal is simple: private, low-cost inference, powered by idle Macs. We’re starting with Gemma 4 and GPT-OSS, then slowly ramping up to larger models as we load test and scale over the next 2 weeks. Providers: go to darkbloom [dot] dev, scroll to down, and run the install command. Thank you to everyone who has been running nodes, giving feedback, and helping us build this network. Waking up the world’s sleeping compute!
4:25

Darkbloom just completed a major network upgrade! BIG UPDATE: We’re moving from Research Preview to Public Alpha. In the last month: - 1000s of provider signups, 250 live providers at peak - 600M+ tokens served With this upgrade, performance is up 30–200% across key metrics like TTFT, TPS, and total tokens served. The goal is simple: private, low-cost inference, powered by idle Macs. We’re starting with Gemma 4 and GPT-OSS, then slowly ramping up to larger models as we load test and scale over the next 2 weeks. Providers: go to darkbloom [dot] dev, scroll to down, and run the install command. Thank you to everyone who has been running nodes, giving feedback, and helping us build this network. Waking up the world’s sleeping compute!

Gajesh

65,997 görüntüleme • 2 ay önce

🎉The Version 1.6 Dev Channel Video Is Coming! 💙Governors, with Version 1.6 coming up, our Dev channel has been updated! This time, we are bringing you detailed information about the whole new KvK playstyle, Golden Expedition, and the much-anticipated migration event. 🎊There's also a surprise gift waiting for you at the end of the video. Don't miss out! 🥰Thank you for your continuous support for our Dev channel! Feel free to leave your suggestions in the comments below. Your valuable words always motivate us to move forward! #AoEMobile #AgeofEmpiresMobile
5:24

🎉The Version 1.6 Dev Channel Video Is Coming! 💙Governors, with Version 1.6 coming up, our Dev channel has been updated! This time, we are bringing you detailed information about the whole new KvK playstyle, Golden Expedition, and the much-anticipated migration event. 🎊There's also a surprise gift waiting for you at the end of the video. Don't miss out! 🥰Thank you for your continuous support for our Dev channel! Feel free to leave your suggestions in the comments below. Your valuable words always motivate us to move forward! #AoEMobile #AgeofEmpiresMobile

Age of Empires Mobile Official

27,062 görüntüleme • 1 yıl önce

All the updates brought by Samsung One UI 9 combined may not be as many as what Home Up has delivered this time. The latest version of Home Up 18.0.00.24 has added the ability to adjust the background color of the Edge Panel. This is my “homework” setup, feel free to use it as a reference. By properly adjusting the blur and transparency levels, you can create a more refined and beautiful visual effect for the Edge Panel. Check out my video to see the results.
0:57

All the updates brought by Samsung One UI 9 combined may not be as many as what Home Up has delivered this time. The latest version of Home Up 18.0.00.24 has added the ability to adjust the background color of the Edge Panel. This is my “homework” setup, feel free to use it as a reference. By properly adjusting the blur and transparency levels, you can create a more refined and beautiful visual effect for the Edge Panel. Check out my video to see the results.

Ice Universe

48,520 görüntüleme • 1 ay önce

Markets change. Human nature doesn’t. Cliff Asness (Clifford Asness) joins Infinite Loops to discuss the gamification of markets, and why extreme valuations keep returning despite decades of data. We explore how trading apps change behavior, why speculation feels productive, and what gets lost when investing becomes entertainment. If you care about long-term thinking in a short-term world, this episode is for you. TIMESTAMPS: 0:00 Intro 1:10 Losses Hurt More than the Wins 6:20 Dot-Com Bubble vs. the 2020 Tech Bubble 13:30 Meme Stocks, Robinhood, and Gamified Investing 21:40 Asness Doesn’t Know Which Stocks He Owns 29:10 Quant Discipline, Models, and Looking Stupid 38:30 Market Crisesand Volatility 52:10 Human Nature as the Last Investing Edge 1:04:40 AI and “Surrendering to the Machines” 1:23:50 Forecasting Bubbles 1:37:40 Private Equity and Volatility Laundering 1:55:30 What Asness Would Change If He Ran the World
1:57:55

Markets change. Human nature doesn’t. Cliff Asness (Clifford Asness) joins Infinite Loops to discuss the gamification of markets, and why extreme valuations keep returning despite decades of data. We explore how trading apps change behavior, why speculation feels productive, and what gets lost when investing becomes entertainment. If you care about long-term thinking in a short-term world, this episode is for you. TIMESTAMPS: 0:00 Intro 1:10 Losses Hurt More than the Wins 6:20 Dot-Com Bubble vs. the 2020 Tech Bubble 13:30 Meme Stocks, Robinhood, and Gamified Investing 21:40 Asness Doesn’t Know Which Stocks He Owns 29:10 Quant Discipline, Models, and Looking Stupid 38:30 Market Crisesand Volatility 52:10 Human Nature as the Last Investing Edge 1:04:40 AI and “Surrendering to the Machines” 1:23:50 Forecasting Bubbles 1:37:40 Private Equity and Volatility Laundering 1:55:30 What Asness Would Change If He Ran the World

Infinite Loops 🎙

184,180 görüntüleme • 6 ay önce

How does Phoenix GenAI running on decentralized AI elastic compute stack up against mainstream Web2 AI platforms? Below is a side-by-side comparison between Kling AI image-to-video vs Phoenix GenAI Img2Motion V2 Beta. Average output time puts Phoenix GenAI Img2Motion V2 Beta at 7-10 min running on SkyNet’s High Performance GPU Cluster, and Kling average output at around 15-30 min. In terms of motion continuity, overall Kling has the advantage, but in terms of aesthetic and picture clarity, Phoenix GenAI seems to have an edge. We’ll let the you be the judge! $PHB #Phoenix #GenAI
0:50

How does Phoenix GenAI running on decentralized AI elastic compute stack up against mainstream Web2 AI platforms? Below is a side-by-side comparison between Kling AI image-to-video vs Phoenix GenAI Img2Motion V2 Beta. Average output time puts Phoenix GenAI Img2Motion V2 Beta at 7-10 min running on SkyNet’s High Performance GPU Cluster, and Kling average output at around 15-30 min. In terms of motion continuity, overall Kling has the advantage, but in terms of aesthetic and picture clarity, Phoenix GenAI seems to have an edge. We’ll let the you be the judge! $PHB #Phoenix #GenAI

Phoenix AI Labs

24,127 görüntüleme • 1 yıl önce

If your MCP server has dozens of tools, it’s probably built wrong. You need tools that are specific and clear for each use case—but you also can’t have too many. This creates an almost impossible tradeoff that most companies don’t know how to solve. That’s why I interviewed my friend Alex Rattray (Alex Rattray), the founder and CEO of Stainless. Stainless builds APIs, SDKs, and MCP servers for companies like OpenAI and Anthropic. Alex has spent years mastering how to make software talk to software, and he came on the show to share what he knows. I had him on Every 📧’s AI & I to talk about MCP and the future of the AI-native internet. We get into: • Design MCP servers to be lean and precise. Alex’s best practices for building reliable MCP servers start with keeping the toolset small, giving each tool a precise name and description, and minimizing the inputs and outputs the model has to handle. At Stainless, they also often add a JSON filter on top to strip out unnecessary data. • Make complex APIs manageable with dynamic mode. To solve the problem of how an AI figures out which tool to use in larger APIs, Stainless switches to “dynamic mode,” where the model gets only three tools: List the endpoints, pick one and learn about it, and then execute it. • MCP servers as business copilots. At Stainless, Alex uses MCP servers to connect tools like Notion and HubSpot, so he can ask questions like, “Which customers signed up last week?” The system queries multiple databases and returns a summary that would’ve otherwise taken multiple logins and searches. • Create a “brain” for your company with Claude Code. Alex built a shared company brain at Stainless by keeping Claude Code running on his system and asking it to save useful inputs—like customer feedback and SQL queries—into GitHub. Over time, this creates a curated archive his team can query easily. • The future of MCP is code execution. Instead of giving models hundreds of tools, Alex believes the most powerful setup will be a simple code execution tool and a doc search tool. The AI writes code against an API’s SDK, runs it on a server, and checks the docs when it gets stuck. This is a must-watch for anyone who wants to understand MCP—and learn how to use them as a competitive edge. Watch below! Timestamps: Introduction: 00:01:14 Why Alex likes running barefoot: 00:02:54 APIs and MCP, the connectors of the new internet: 00:05:09 Why MCP servers are hard to get right: 00:10:53 Design principles for reliable MCP servers: 00:20:07 Scaling MCP servers for large APIs: 00:23:50 Using MCP for business ops at Stainless: 00:25:14 Building a company brain with Claude Code: 00:28:12 Where MCP goes from here: 00:33:59 Alex’s take on the security model for MCP: 00:41:10
51:39

If your MCP server has dozens of tools, it’s probably built wrong. You need tools that are specific and clear for each use case—but you also can’t have too many. This creates an almost impossible tradeoff that most companies don’t know how to solve. That’s why I interviewed my friend Alex Rattray (Alex Rattray), the founder and CEO of Stainless. Stainless builds APIs, SDKs, and MCP servers for companies like OpenAI and Anthropic. Alex has spent years mastering how to make software talk to software, and he came on the show to share what he knows. I had him on Every 📧’s AI & I to talk about MCP and the future of the AI-native internet. We get into: • Design MCP servers to be lean and precise. Alex’s best practices for building reliable MCP servers start with keeping the toolset small, giving each tool a precise name and description, and minimizing the inputs and outputs the model has to handle. At Stainless, they also often add a JSON filter on top to strip out unnecessary data. • Make complex APIs manageable with dynamic mode. To solve the problem of how an AI figures out which tool to use in larger APIs, Stainless switches to “dynamic mode,” where the model gets only three tools: List the endpoints, pick one and learn about it, and then execute it. • MCP servers as business copilots. At Stainless, Alex uses MCP servers to connect tools like Notion and HubSpot, so he can ask questions like, “Which customers signed up last week?” The system queries multiple databases and returns a summary that would’ve otherwise taken multiple logins and searches. • Create a “brain” for your company with Claude Code. Alex built a shared company brain at Stainless by keeping Claude Code running on his system and asking it to save useful inputs—like customer feedback and SQL queries—into GitHub. Over time, this creates a curated archive his team can query easily. • The future of MCP is code execution. Instead of giving models hundreds of tools, Alex believes the most powerful setup will be a simple code execution tool and a doc search tool. The AI writes code against an API’s SDK, runs it on a server, and checks the docs when it gets stuck. This is a must-watch for anyone who wants to understand MCP—and learn how to use them as a competitive edge. Watch below! Timestamps: Introduction: 00:01:14 Why Alex likes running barefoot: 00:02:54 APIs and MCP, the connectors of the new internet: 00:05:09 Why MCP servers are hard to get right: 00:10:53 Design principles for reliable MCP servers: 00:20:07 Scaling MCP servers for large APIs: 00:23:50 Using MCP for business ops at Stainless: 00:25:14 Building a company brain with Claude Code: 00:28:12 Where MCP goes from here: 00:33:59 Alex’s take on the security model for MCP: 00:41:10

Dan Shipper 📧

15,645 görüntüleme • 10 ay önce

Tech CEO Amjad Masad on the cults of Silicon Valley. This is the deepest and most interesting explanation of AI you’ll ever see. (0:00) Artificial Intelligence (10:01) Bitcoin (22:30) The Extropians Cult (29:15) Transhumanism (34:04) Longtermism (40:24) Are Machines Capable of Thinking? (43:11) The Two Conflicting Theories of Physics (49:46) The Difference Between Mind and Computer (1:04:24) The Abuse of Technology (1:12:36) Virtual Companionship (1:15:12) Do Silicon Valley Elites Acknowledge the Existence of God? (1:26:37) Silicon Valley Turning to Donald Trump (1:40:20) Elon Musk and Free Speech Includes paid partnerships.
1:52:53

Tech CEO Amjad Masad on the cults of Silicon Valley. This is the deepest and most interesting explanation of AI you’ll ever see. (0:00) Artificial Intelligence (10:01) Bitcoin (22:30) The Extropians Cult (29:15) Transhumanism (34:04) Longtermism (40:24) Are Machines Capable of Thinking? (43:11) The Two Conflicting Theories of Physics (49:46) The Difference Between Mind and Computer (1:04:24) The Abuse of Technology (1:12:36) Virtual Companionship (1:15:12) Do Silicon Valley Elites Acknowledge the Existence of God? (1:26:37) Silicon Valley Turning to Donald Trump (1:40:20) Elon Musk and Free Speech Includes paid partnerships.

Tucker Carlson

3,794,562 görüntüleme • 2 yıl önce

AI video is moving from random generation to real creative control. Seedance Enterprise 2.5 on BytePlus Lumina delivers ultra realistic, cinematic quality with lifelike physical lighting and textures. By combining Seedream visuals, SeedAudio references, and up to 50 multimodal inputs, you can create fully cohesive videos up to 30 seconds that look and feel like real footage. You can stage scenes, control camera movement, position characters, and build more consistent narratives with the 3D Director’s Desk. Less guessing. More directing. Directed by me. Powered by BytePlus Lumina. Seedance 2.5 API is about to launch on BytePlus, and now you can try it directly on Lumina.
0:30

AI video is moving from random generation to real creative control. Seedance Enterprise 2.5 on BytePlus Lumina delivers ultra realistic, cinematic quality with lifelike physical lighting and textures. By combining Seedream visuals, SeedAudio references, and up to 50 multimodal inputs, you can create fully cohesive videos up to 30 seconds that look and feel like real footage. You can stage scenes, control camera movement, position characters, and build more consistent narratives with the 3D Director’s Desk. Less guessing. More directing. Directed by me. Powered by BytePlus Lumina. Seedance 2.5 API is about to launch on BytePlus, and now you can try it directly on Lumina.

Doreen

82,963 görüntüleme • 6 gün önce

AI has transformed how video is created. We think the next wave is about understanding it. Over the past few years, we've seen remarkable advances in video generation, editing, avatars, and creative tooling. An increasingly important problem is teaching machines to search, analyze, reason over, and extract insight from video - across massive libraries and live streams alike. We're calling this video intelligence, and we're actively looking to back founders building here. We're most excited about companies pushing on the core capabilities: - Video-native models - multimodal embeddings, temporal reasoning, and retrieval built specifically for video rather than adapted from image or text - Real-time and large-scale pipelines - infrastructure for processing, indexing, and querying video at the speed and scale enterprises actually need - Agentic and reasoning layers - systems that don't just retrieve clips but answer questions, surface anomalies, and take action on what they see The models and infrastructure to make this real are appearing to be crossing a capability threshold right now. Multimodal foundation models are maturing, storage costs have collapsed, and enterprises are sitting on years of unstructured video with no way to use it. That infrastructure unlocks a wide range of applications including media and sports workflows, security and physical operations, enterprise knowledge management, advertising analytics, robotics, and consumer products, where video has historically been dark data. If you're building in video intelligence at the model layer, the platform layer, or in a vertical application, we'd love to talk!
1:41

AI has transformed how video is created. We think the next wave is about understanding it. Over the past few years, we've seen remarkable advances in video generation, editing, avatars, and creative tooling. An increasingly important problem is teaching machines to search, analyze, reason over, and extract insight from video - across massive libraries and live streams alike. We're calling this video intelligence, and we're actively looking to back founders building here. We're most excited about companies pushing on the core capabilities: - Video-native models - multimodal embeddings, temporal reasoning, and retrieval built specifically for video rather than adapted from image or text - Real-time and large-scale pipelines - infrastructure for processing, indexing, and querying video at the speed and scale enterprises actually need - Agentic and reasoning layers - systems that don't just retrieve clips but answer questions, surface anomalies, and take action on what they see The models and infrastructure to make this real are appearing to be crossing a capability threshold right now. Multimodal foundation models are maturing, storage costs have collapsed, and enterprises are sitting on years of unstructured video with no way to use it. That infrastructure unlocks a wide range of applications including media and sports workflows, security and physical operations, enterprise knowledge management, advertising analytics, robotics, and consumer products, where video has historically been dark data. If you're building in video intelligence at the model layer, the platform layer, or in a vertical application, we'd love to talk!

Jason Cui

36,205 görüntüleme • 3 ay önce

nemotron 3 omni q8 on dgx spark 128gb vram cranking via hermes agent at 56 tok/s. first night of real local agentic on this box and local hits harder than i thought it would. q8 (near lossless quant, perplexity loss <1% vs fp16) running 256k context on 33 gb of unified memory, 90+ gb still free. multimodal omni weights included. hermes agent driving from telegram, talking to it from bed. speed: 56 tok/s generation, 1,300 tok/s prefill. for context, qwen 3.6 27b at q4 (heavy quant) on 3090 = 40 tok/s. nemotron at higher precision quant on spark beats qwen at lower precision quant on 3090. moe 3.5b active params architecture earns its keep. what i tested tonight: agentic tool calling works clean. ask it to check disks, it autonomously runs df -h through hermes agent. ask it to set up telegram gateway, it invokes the hermes-agent skill, walks through the prompts, completes the flow. overthinks a bit before tool calls (reasoning model trait) but lands the right move every time. researches api docs, internalizes, tests, documents. completes tasks. current models on dgx spark: 9 gguf files, 305 gb total, mix of qwen 3.6 27b dense (5 quants), nemotron omni (4 quants), deepseek v4-flash 158b q4 (the 112gb flagship test). more data coming this week as i benchmark each.
1:40

nemotron 3 omni q8 on dgx spark 128gb vram cranking via hermes agent at 56 tok/s. first night of real local agentic on this box and local hits harder than i thought it would. q8 (near lossless quant, perplexity loss <1% vs fp16) running 256k context on 33 gb of unified memory, 90+ gb still free. multimodal omni weights included. hermes agent driving from telegram, talking to it from bed. speed: 56 tok/s generation, 1,300 tok/s prefill. for context, qwen 3.6 27b at q4 (heavy quant) on 3090 = 40 tok/s. nemotron at higher precision quant on spark beats qwen at lower precision quant on 3090. moe 3.5b active params architecture earns its keep. what i tested tonight: agentic tool calling works clean. ask it to check disks, it autonomously runs df -h through hermes agent. ask it to set up telegram gateway, it invokes the hermes-agent skill, walks through the prompts, completes the flow. overthinks a bit before tool calls (reasoning model trait) but lands the right move every time. researches api docs, internalizes, tests, documents. completes tasks. current models on dgx spark: 9 gguf files, 305 gb total, mix of qwen 3.6 27b dense (5 quants), nemotron omni (4 quants), deepseek v4-flash 158b q4 (the 112gb flagship test). more data coming this week as i benchmark each.

Sudo su

30,228 görüntüleme • 3 ay önce