Загрузка видео...

Не удалось загрузить видео

На главную

Today, we're announcing our acquisition of FlashDocs, the leader in generative slide deck creation. FlashDocs automatically turns LLM output into fully branded slide decks, creating presentations in seconds, not hours.

hebbia's profile picture

Hebbia

15,108 subscribers

424,539 просмотров • 1 год назад •via X (Twitter)

Наука и технологии
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

I really like how it takes some of the slides which are static (each one of these is its own slide in the original deck) and turns them into animations.
0:11

I really like how it takes some of the slides which are static (each one of these is its own slide in the original deck) and turns them into animations.

Thariq

14,965 просмотров • 1 месяц назад

Onboarding the next 100M gamers into Web3 requires 100x lower DA fees. Announcing the upgrade of Ancient8 Chain with Celestia underneath ✨, the leader in modular data availability. (🎮,🟢) Gaming developers can now build fully on chain experiences and onboard new waves of users not possible before, unconstrained by throughput. Read more here 👉🏻
0:12

Onboarding the next 100M gamers into Web3 requires 100x lower DA fees. Announcing the upgrade of Ancient8 Chain with Celestia underneath ✨, the leader in modular data availability. (🎮,🟢) Gaming developers can now build fully on chain experiences and onboard new waves of users not possible before, unconstrained by throughput. Read more here 👉🏻

Ancient8

421,761 просмотров • 2 лет назад

September was packed with upgrades at Hebbia, including powerful workflow automations that let you build expert-level financial models, craft polished presentations, and more. This edition of “The Disclosure” is live now.
0:14

September was packed with upgrades at Hebbia, including powerful workflow automations that let you build expert-level financial models, craft polished presentations, and more. This edition of “The Disclosure” is live now.

Hebbia

755,846 просмотров • 10 месяцев назад

Now available: Dark Mode in Hebbia. Built for long hours and deep focus – in a design that adapts to how investors and bankers actually work.
0:13

Now available: Dark Mode in Hebbia. Built for long hours and deep focus – in a design that adapts to how investors and bankers actually work.

Hebbia

1,028,982 просмотров • 10 месяцев назад

🌟2025 is looking bright! Today, we announced that AlphaSense has surpassed $400M in ARR and is now bringing our cutting-edge AI search capabilities to over 6,000 businesses. It has been an incredible year of milestones, from our acquisition of Tegus by AlphaSense and increasing our valuation to $4B, to being named a CNBC Disruptor 50 company. We are excited to continue introducing some of our biggest product innovations to the platform this year – the groundbreaking additions of Generative Search and Generative Grid, and our latest partnership with , are just the beginning of what’s to come. This growth would not be possible without our customers. Thank you for trusting AlphaSense to gain deeper insights, make data-driven decisions, and stay ahead of market trends with speed. Read the full announcement:
0:17

🌟2025 is looking bright! Today, we announced that AlphaSense has surpassed $400M in ARR and is now bringing our cutting-edge AI search capabilities to over 6,000 businesses. It has been an incredible year of milestones, from our acquisition of Tegus by AlphaSense and increasing our valuation to $4B, to being named a CNBC Disruptor 50 company. We are excited to continue introducing some of our biggest product innovations to the platform this year – the groundbreaking additions of Generative Search and Generative Grid, and our latest partnership with , are just the beginning of what’s to come. This growth would not be possible without our customers. Thank you for trusting AlphaSense to gain deeper insights, make data-driven decisions, and stay ahead of market trends with speed. Read the full announcement:

AlphaSense

30,758 просмотров • 1 год назад

ByteDance just open sourced an AI SuperAgent that can research, code, build websites, create slide decks, and generate videos. All by itself. DeerFlow 2.0 (27K+ GitHub stars ⭐️), an AI system acting like an autonomous employee with its own computer workspace to research and code. Standard chatbots only generate text and forget your preferences. DeerFlow solves this by giving the AI an isolated virtual computer environment where it safely runs programs. When given a massive task, the main program creates several smaller AI assistants to work simultaneously. It also saves your past workflows so it gets smarter about your needs. DeerFlow is model-agnostic — it works with any LLM that implements the OpenAI-compatible API. Fully supports running local models on your own computer using tools like Ollama. An example - you ask for research on the top 10 AI startups in 2026 for a presentation, the lead agent in DeerFlow breaks that big job into smaller sub-tasks. It assigns one sub-agent to look into each company, another to find funding details, and a third to handle competitor analysis. These agents do all their work in parallel. Everything eventually converges, and a final agent pulls the results into a slide deck complete with custom visuals.
0:17

ByteDance just open sourced an AI SuperAgent that can research, code, build websites, create slide decks, and generate videos. All by itself. DeerFlow 2.0 (27K+ GitHub stars ⭐️), an AI system acting like an autonomous employee with its own computer workspace to research and code. Standard chatbots only generate text and forget your preferences. DeerFlow solves this by giving the AI an isolated virtual computer environment where it safely runs programs. When given a massive task, the main program creates several smaller AI assistants to work simultaneously. It also saves your past workflows so it gets smarter about your needs. DeerFlow is model-agnostic — it works with any LLM that implements the OpenAI-compatible API. Fully supports running local models on your own computer using tools like Ollama. An example - you ask for research on the top 10 AI startups in 2026 for a presentation, the lead agent in DeerFlow breaks that big job into smaller sub-tasks. It assigns one sub-agent to look into each company, another to find funding details, and a third to handle competitor analysis. These agents do all their work in parallel. Everything eventually converges, and a final agent pulls the results into a slide deck complete with custom visuals.

Rohan Paul

50,097 просмотров • 5 месяцев назад

Yesterday, I was staring at a 100-page PDF for my project review… fully convinced my weekend was officially over. My professor wanted a clean consulting-style deck. I had zero slides. Zero energy. And a very real fear of PowerPoint. Out of panic, I dropped the PDF into Kimi.ai, expecting nothing. But in seconds, it turned the whole document into a neat, structured deck. Not random slides but actual infographics, charts, and visual storytelling that looked far better than anything I could design at 2 AM. What helped the most was how it handled the messy parts. It cleaned up my unorganized notes, rebuilt them into clear slides, and still exported everything as a fully editable PPTX, which meant I could tweak every detail exactly how my professor wanted. This tool did what Nano banana can’t. It makes everything unlimited and fully editable which honestly saved me hours.
0:11

Yesterday, I was staring at a 100-page PDF for my project review… fully convinced my weekend was officially over. My professor wanted a clean consulting-style deck. I had zero slides. Zero energy. And a very real fear of PowerPoint. Out of panic, I dropped the PDF into Kimi.ai, expecting nothing. But in seconds, it turned the whole document into a neat, structured deck. Not random slides but actual infographics, charts, and visual storytelling that looked far better than anything I could design at 2 AM. What helped the most was how it handled the messy parts. It cleaned up my unorganized notes, rebuilt them into clear slides, and still exported everything as a fully editable PPTX, which meant I could tweak every detail exactly how my professor wanted. This tool did what Nano banana can’t. It makes everything unlimited and fully editable which honestly saved me hours.

siezzie

52,269 просмотров • 8 месяцев назад

LLM Knowledge Base → Slides When Andrej Karpathy shared his LLM Knowledge Base setup, many were wondering how to generate more visual forms of the wiki. There are many options, but I think Gamma is one of the best at producing high-quality, rich presentations. To showcase this, I just built a pipeline that turns my AI papers wiki (1K+ papers across 20 AI agent topics) into polished slide presentations using Gamma. The flow: Obsidian vault → Gamma MCP → embedded preview in my dashboard. I give one command to my agent, which pulls the top papers from each topic (via the wiki), feeds them to Gamma, and renders the presentation inline. The Gamma connector for Claude is a great choice for generating beautiful and professional slides. Easy to use. Go to your Claude instance and add the official Gamma connector. That's it! Claude Code will now have access to all the necessary MCP tools for generating slides. I use the Claude Agent SDK for my agent orchestrator, so I use the official Gamma MCP tools and embed the generated slides in an iframe via my artifact preview. See the clip below for an example.
0:19

LLM Knowledge Base → Slides When Andrej Karpathy shared his LLM Knowledge Base setup, many were wondering how to generate more visual forms of the wiki. There are many options, but I think Gamma is one of the best at producing high-quality, rich presentations. To showcase this, I just built a pipeline that turns my AI papers wiki (1K+ papers across 20 AI agent topics) into polished slide presentations using Gamma. The flow: Obsidian vault → Gamma MCP → embedded preview in my dashboard. I give one command to my agent, which pulls the top papers from each topic (via the wiki), feeds them to Gamma, and renders the presentation inline. The Gamma connector for Claude is a great choice for generating beautiful and professional slides. Easy to use. Go to your Claude instance and add the official Gamma connector. That's it! Claude Code will now have access to all the necessary MCP tools for generating slides. I use the Claude Agent SDK for my agent orchestrator, so I use the official Gamma MCP tools and embed the generated slides in an iframe via my artifact preview. See the clip below for an example.

elvis

47,966 просмотров • 4 месяцев назад

Gamma API + Claude + n8n is absolutely WILD This combo is absolutely insane for turning meetings into professional decks automatically. No manual notes. No "I'll send that over later." No remembering who needs what. Just AI tools working together like a professional operations team. Here's how it works: → Meeting ends, n8n trigger fires and pulls Fireflies transcript → Claude analyzes everything and creates professional presentation structure → Gamma API generates designer-quality deck with interactive link + PDF → Slack sends you preview with Approve/Reject buttons → Hit Approve, system emails deck to ALL participants with action items → They get PDF attachment + Gamma link before you even close Zoom Perfect for founders and sales teams who want to look impossibly organized. The power is in the combo: n8n = zero-click automation that runs itself Claude = extracts what matters, structures it professionally Gamma = designer-quality decks that look like you hired a team While others are scrambling to remember action items, your deck is already in their inbox. Close deals faster. Look impossibly organized. Never miss follow-up again. Like, RT + reply with "GAMMA" and I'll DM the complete system (Must be following so I can DM) Skip this and keep manually building slide decks at 11pm.
0:14

Gamma API + Claude + n8n is absolutely WILD This combo is absolutely insane for turning meetings into professional decks automatically. No manual notes. No "I'll send that over later." No remembering who needs what. Just AI tools working together like a professional operations team. Here's how it works: → Meeting ends, n8n trigger fires and pulls Fireflies transcript → Claude analyzes everything and creates professional presentation structure → Gamma API generates designer-quality deck with interactive link + PDF → Slack sends you preview with Approve/Reject buttons → Hit Approve, system emails deck to ALL participants with action items → They get PDF attachment + Gamma link before you even close Zoom Perfect for founders and sales teams who want to look impossibly organized. The power is in the combo: n8n = zero-click automation that runs itself Claude = extracts what matters, structures it professionally Gamma = designer-quality decks that look like you hired a team While others are scrambling to remember action items, your deck is already in their inbox. Close deals faster. Look impossibly organized. Never miss follow-up again. Like, RT + reply with "GAMMA" and I'll DM the complete system (Must be following so I can DM) Skip this and keep manually building slide decks at 11pm.

Samruddhi Mokal

10,795 просмотров • 10 месяцев назад

Today we’re open-sourcing Stable Audio Open Small, a 341M-parameter text-to-audio model optimized to run entirely on Arm CPUs. This means 99% of smartphones can now generate music-production samples in seconds, right on-device with no internet required. Built for fast, on-the-go creation, it turns your next quick idea into up to 11 seconds of audio. Generate drum loops, foley, riffs, and textures right where you are. No cords 🔌 just chords 🎹 You can learn more here:
0:17

Today we’re open-sourcing Stable Audio Open Small, a 341M-parameter text-to-audio model optimized to run entirely on Arm CPUs. This means 99% of smartphones can now generate music-production samples in seconds, right on-device with no internet required. Built for fast, on-the-go creation, it turns your next quick idea into up to 11 seconds of audio. Generate drum loops, foley, riffs, and textures right where you are. No cords 🔌 just chords 🎹 You can learn more here:

Stability AI

94,796 просмотров • 1 год назад

i analysed 1,000 TIKTOK slideshows for consumer apps... here's what i found something that change how you run your app/saas campaign since. most people assume the slideshow with the most views brings in the most installs. i tracked every metric i could pull across 1000+ posts. views, saves, comment sentiment, slide count, where the app got mentioned in the sequence, caption length, niche. the data told a different story. the highest converting slideshows rarely broke 100k views. some sat under 20k. meanwhile some of the viral ones with 2m+ views converted under 0.05%. viral and profitable are two different games. here's the pattern that separated the winners. the format was almost always: content, content, content, content, ad warmup, app push. 4 to 6 slides that feel like normal lifestyle or niche content, no mention of the app at all. then one slide near the end where the product shows up, framed as part of the story instead of an ad. apps that opened with the product on slide 1 underperformed almost every time. the accounts winning were disguising the app inside content people were already scrolling for anyway. travel aesthetics, interior inspiration, "things nobody tells you about x" hooks, niche opinions. the second pattern was volume, not virality. accounts running 8-10 tiktoks, posting 2x a day, same slide formats with fresh variations each time. 90% of posts stayed under 5k views, most even under 300. a handful hit 50k-500k. a rare one crossed 1m. the accounts winning weren't making one perfect post, they were running the format enough times that the algorithm found the winners for them. the workflow behind it, if you want to copy it: research first. search your niche on tiktok, screenshot every top slideshow, save the captions somewhere. this is the raw material for everything after. pull matching visuals from pinterest for each slide type in that screenshot pile. figure out what slide 1 looks like, slide 2, slide 3. download a handful per slide type. not everything from pinterest can be reposted as is. run it through an image api like openai or gemini to generate variations that keep the same vibe. 5 slide types x 100 variations gets you 500 usable images fast. feed the competitor captions into claude code and have it write new caption variations in that same tone, keep the early slides content only, drop the app in near the end. claude code can overlay the captions onto the images directly using ffmpeg, then hand scheduling off to a tool that allows accounts to post automatically without you touching them daily. set this up once and you get months of content queued across every account, running the same proven format with fresh visuals each time. the accounts losing were treating every post as a one off. the accounts winning built a system and let volume do the work.
0:18

i analysed 1,000 TIKTOK slideshows for consumer apps... here's what i found something that change how you run your app/saas campaign since. most people assume the slideshow with the most views brings in the most installs. i tracked every metric i could pull across 1000+ posts. views, saves, comment sentiment, slide count, where the app got mentioned in the sequence, caption length, niche. the data told a different story. the highest converting slideshows rarely broke 100k views. some sat under 20k. meanwhile some of the viral ones with 2m+ views converted under 0.05%. viral and profitable are two different games. here's the pattern that separated the winners. the format was almost always: content, content, content, content, ad warmup, app push. 4 to 6 slides that feel like normal lifestyle or niche content, no mention of the app at all. then one slide near the end where the product shows up, framed as part of the story instead of an ad. apps that opened with the product on slide 1 underperformed almost every time. the accounts winning were disguising the app inside content people were already scrolling for anyway. travel aesthetics, interior inspiration, "things nobody tells you about x" hooks, niche opinions. the second pattern was volume, not virality. accounts running 8-10 tiktoks, posting 2x a day, same slide formats with fresh variations each time. 90% of posts stayed under 5k views, most even under 300. a handful hit 50k-500k. a rare one crossed 1m. the accounts winning weren't making one perfect post, they were running the format enough times that the algorithm found the winners for them. the workflow behind it, if you want to copy it: research first. search your niche on tiktok, screenshot every top slideshow, save the captions somewhere. this is the raw material for everything after. pull matching visuals from pinterest for each slide type in that screenshot pile. figure out what slide 1 looks like, slide 2, slide 3. download a handful per slide type. not everything from pinterest can be reposted as is. run it through an image api like openai or gemini to generate variations that keep the same vibe. 5 slide types x 100 variations gets you 500 usable images fast. feed the competitor captions into claude code and have it write new caption variations in that same tone, keep the early slides content only, drop the app in near the end. claude code can overlay the captions onto the images directly using ffmpeg, then hand scheduling off to a tool that allows accounts to post automatically without you touching them daily. set this up once and you get months of content queued across every account, running the same proven format with fresh visuals each time. the accounts losing were treating every post as a one off. the accounts winning built a system and let volume do the work.

Mufasa

41,947 просмотров • 19 дней назад

We're premiering our launchpad tech. You get spotlight featured, custom UI, fulfilment support + email blast to one of the largest collector bases in the world. We've locked in some incredible IPs across web3 and web2 all dropping over the next few months. Markets are dumping yet CT is having fun + supporting Wut and the fwog IP. Without creators like him, there would be no color and fun on the timeline. Indulge in community, support one another, and forget about short term macro. This space is so much more than price action. And if you're building something cool, slide into phygitals DMs, you'll get unbeatable dev + infra support :)
0:19

We're premiering our launchpad tech. You get spotlight featured, custom UI, fulfilment support + email blast to one of the largest collector bases in the world. We've locked in some incredible IPs across web3 and web2 all dropping over the next few months. Markets are dumping yet CT is having fun + supporting Wut and the fwog IP. Without creators like him, there would be no color and fun on the timeline. Indulge in community, support one another, and forget about short term macro. This space is so much more than price action. And if you're building something cool, slide into phygitals DMs, you'll get unbeatable dev + infra support :)

Jake | Phygitals

27,725 просмотров • 9 месяцев назад

I tried MiniMax Design (H3) to see how it handles real content creation. The workflow is simple. You just write a prompt or drop in an image, and it turns that into a dynamic video with motion, framing, and scene depth. No timeline to manage. No editing setup. No back and forth. What stood out to me: • Text to video and image to video both feel smooth. • It handles motion, camera angles, and flow on its own. • Output is fast, usually within seconds. • Works well for reels, quick ads, storytelling, and idea testing. It removes the hardest part: starting from scratch and turns your ideas into content in minutes. Instead of thinking, “How do I make this video?” You start with, “What do I want to create?” That shift alone makes it worth exploring. Try it here: #Hailuo
0:10

I tried MiniMax Design (H3) to see how it handles real content creation. The workflow is simple. You just write a prompt or drop in an image, and it turns that into a dynamic video with motion, framing, and scene depth. No timeline to manage. No editing setup. No back and forth. What stood out to me: • Text to video and image to video both feel smooth. • It handles motion, camera angles, and flow on its own. • Output is fast, usually within seconds. • Works well for reels, quick ads, storytelling, and idea testing. It removes the hardest part: starting from scratch and turns your ideas into content in minutes. Instead of thinking, “How do I make this video?” You start with, “What do I want to create?” That shift alone makes it worth exploring. Try it here: #Hailuo

Manish Kumar Shah

27,680 просмотров • 4 месяцев назад

✍️Today is truly a day of rejoicing🎊🎉🫑🫑🫑🫑 ✍️Our First Anambra Greenhouse farms Cluster in Awka Anambra State is almost ready ✍️Remember our vision and goal.. ✍️18,690 units of Greenhouses...$243million new agro investments in the South East of Nigeria ✍️Annual revenue of over $430million into the next 20 years ✍️Job creation of over 36,000 direct jobs and 50,000 indirect jobs ✍️Creating massive wealth and food security for our people ✍️All our agro commodities will be produced organically ✍️We are also rolling out an app and logistics company to deliver our agro products , in any city where we operate from farms to your homes , at 50% the price in the open market ..mm Power of vision and drive ...Join us Thank you Ned Nwogu and Austin Ufomba for believing in me Thank you, Ikenna Umezurike and Uchechi Bede Okoro for going on this journey with me Let's go ... Please hand over your gated home and land ..let's do this more .. What's app 09090900989 OBIARAERI, Nnaemeka Onyeka
0:11

✍️Today is truly a day of rejoicing🎊🎉🫑🫑🫑🫑 ✍️Our First Anambra Greenhouse farms Cluster in Awka Anambra State is almost ready ✍️Remember our vision and goal.. ✍️18,690 units of Greenhouses...$243million new agro investments in the South East of Nigeria ✍️Annual revenue of over $430million into the next 20 years ✍️Job creation of over 36,000 direct jobs and 50,000 indirect jobs ✍️Creating massive wealth and food security for our people ✍️All our agro commodities will be produced organically ✍️We are also rolling out an app and logistics company to deliver our agro products , in any city where we operate from farms to your homes , at 50% the price in the open market ..mm Power of vision and drive ...Join us Thank you Ned Nwogu and Austin Ufomba for believing in me Thank you, Ikenna Umezurike and Uchechi Bede Okoro for going on this journey with me Let's go ... Please hand over your gated home and land ..let's do this more .. What's app 09090900989 OBIARAERI, Nnaemeka Onyeka

OBIARAERI, Nnaemeka Onyeka.

24,989 просмотров • 2 лет назад

🚨 JUST IN: CHINA just released an AI EMPLOYEE that works 24X7 on its own. 100% OPEN SOURCE. It researches, codes, builds websites, creates slide decks, and generates videos. All by itself. All on your computer. It's called DeerFlow. You give it a task. It makes a plan, spins up its own team of sub-agents, and gets to work. You come back and there's a finished deliverable waiting. Not a draft. Not a summary. The actual thing. Not a chatbot. Not a research assistant. An AI with its own computer that works while you sleep. Here's what it does on its own: → Spawns multiple sub-agents in parallel, each tackling a different piece of your task, then combines everything into one finished output → Writes real code, runs it, reads the results, and fixes its own mistakes without asking you once → Builds slide decks, websites, full research reports, and data dashboards from scratch → Remembers you across sessions. Your writing style. Your tech stack. Your preferences. Gets better every time. → Reads files you upload, works with them inside its own filesystem, hands you clean finished outputs → Searches the web, runs commands, calls any tool you plug in Here's how it thinks: You give one instruction. The lead agent makes a plan. Sub-agents fan out and work in parallel. Results come back. Everything gets synthesized. You get a deliverable. A single research task might split into a dozen sub-agents, each exploring a different angle, then converge into one finished website with generated visuals. Here's the wildest part: DeerFlow 2.0 launched on February 28th 2026 and hit number 1 on all of GitHub Trending the same day. Version 2.0 was a complete rewrite. Zero shared code with version 1. Because users kept using it for things the team never intended. Data pipelines. Dashboards. Entire content workflows. The community told them what it needed to become. So they burned it down and rebuilt it. 22.7K GitHub stars. 2.7K forks. Built by ByteDance 100% Open Source. MIT License.
0:17

🚨 JUST IN: CHINA just released an AI EMPLOYEE that works 24X7 on its own. 100% OPEN SOURCE. It researches, codes, builds websites, creates slide decks, and generates videos. All by itself. All on your computer. It's called DeerFlow. You give it a task. It makes a plan, spins up its own team of sub-agents, and gets to work. You come back and there's a finished deliverable waiting. Not a draft. Not a summary. The actual thing. Not a chatbot. Not a research assistant. An AI with its own computer that works while you sleep. Here's what it does on its own: → Spawns multiple sub-agents in parallel, each tackling a different piece of your task, then combines everything into one finished output → Writes real code, runs it, reads the results, and fixes its own mistakes without asking you once → Builds slide decks, websites, full research reports, and data dashboards from scratch → Remembers you across sessions. Your writing style. Your tech stack. Your preferences. Gets better every time. → Reads files you upload, works with them inside its own filesystem, hands you clean finished outputs → Searches the web, runs commands, calls any tool you plug in Here's how it thinks: You give one instruction. The lead agent makes a plan. Sub-agents fan out and work in parallel. Results come back. Everything gets synthesized. You get a deliverable. A single research task might split into a dozen sub-agents, each exploring a different angle, then converge into one finished website with generated visuals. Here's the wildest part: DeerFlow 2.0 launched on February 28th 2026 and hit number 1 on all of GitHub Trending the same day. Version 2.0 was a complete rewrite. Zero shared code with version 1. Because users kept using it for things the team never intended. Data pipelines. Dashboards. Entire content workflows. The community told them what it needed to become. So they burned it down and rebuilt it. 22.7K GitHub stars. 2.7K forks. Built by ByteDance 100% Open Source. MIT License.

Kanika

738,256 просмотров • 4 месяцев назад

A HARDWARE MODDER BOLTED A BATTERY PACK ONTO A $599 MAC MINI M4 AND TURNED A DESK SERVER INTO A BACKPACK AI RIG THAT RUNS FOR EIGHT HOURS WITHOUT A WALL OUTLET he posts a video of the mac mini with a slim battery module clamped to the side, USB-C plugged into the back. no power brick, no outlet, just the 10 to 30W draw of the M4 chip pulling from a 20,000mAh bank. a full work day on one charge. this is not a portable laptop trick. it is a desk-class AI server that fits in a sling bag the same $599 mac mini i recommended in last week's article as the silent default, now untethered. pair it with a mobile hotspot or a Starlink mini and you have a fully off-grid LLM stack that answers questions in a forest, on a plane, in a power outage. the model lives in the box, the battery powers the box, that is the whole stack this is not a gadget. it is the first time the words "AI server" and "fits in your bag" belong in the same sentence
0:19

A HARDWARE MODDER BOLTED A BATTERY PACK ONTO A $599 MAC MINI M4 AND TURNED A DESK SERVER INTO A BACKPACK AI RIG THAT RUNS FOR EIGHT HOURS WITHOUT A WALL OUTLET he posts a video of the mac mini with a slim battery module clamped to the side, USB-C plugged into the back. no power brick, no outlet, just the 10 to 30W draw of the M4 chip pulling from a 20,000mAh bank. a full work day on one charge. this is not a portable laptop trick. it is a desk-class AI server that fits in a sling bag the same $599 mac mini i recommended in last week's article as the silent default, now untethered. pair it with a mobile hotspot or a Starlink mini and you have a fully off-grid LLM stack that answers questions in a forest, on a plane, in a power outage. the model lives in the box, the battery powers the box, that is the whole stack this is not a gadget. it is the first time the words "AI server" and "fits in your bag" belong in the same sentence

shmidt

31,619 просмотров • 1 месяц назад

The fuel shortage is pushing "the world's biggest parking lot" into the future Hours-long queues at gas stations and rising fuel prices are driving Russians to switch to electric vehicles. According to Autostat, sales of new EVs and hybrids jumped in June, just as the fuel crisis entered its most acute phase. Some owners say their commuting costs have dropped by around 80%. It turns out that Russia is being pushed toward modern technology not by progress, but by shortages. Meanwhile, those who still can't afford an electric car are already getting a "greeting from the future": "In this future, 40 liters of AI-95 gasoline cost 10,000 rubles. Here in Crimea, we're already living in the future."
0:11

The fuel shortage is pushing "the world's biggest parking lot" into the future Hours-long queues at gas stations and rising fuel prices are driving Russians to switch to electric vehicles. According to Autostat, sales of new EVs and hybrids jumped in June, just as the fuel crisis entered its most acute phase. Some owners say their commuting costs have dropped by around 80%. It turns out that Russia is being pushed toward modern technology not by progress, but by shortages. Meanwhile, those who still can't afford an electric car are already getting a "greeting from the future": "In this future, 40 liters of AI-95 gasoline cost 10,000 rubles. Here in Crimea, we're already living in the future."

NEXTA

64,567 просмотров • 1 месяц назад

We removed 93.25% of the connections in our Un-0 image model, fully expecting to pay for it in quality. But it got better. FID 7.15 on ImageNet 64x64, roughly 1.9 ahead of the dense baseline at matched size. Same family of model, a fraction of the couplings, a better score. Here is why that is not as strange as it sounds. Un-0 is a coupled oscillator model, and in the dense version every oscillator talks to every other one. That sounds like a strength, but it means the whole system can fall into catastrophic synchronization: everything locks into step, gradients go flat, and learning stalls. Sparser connectivity leaves room for coherent and incoherent activity to coexist. The dynamics stay alive, and the model keeps learning. Connectivity turns out to be a control knob, not a dial you turn up until it stops. Learn more here:
0:12

We removed 93.25% of the connections in our Un-0 image model, fully expecting to pay for it in quality. But it got better. FID 7.15 on ImageNet 64x64, roughly 1.9 ahead of the dense baseline at matched size. Same family of model, a fraction of the couplings, a better score. Here is why that is not as strange as it sounds. Un-0 is a coupled oscillator model, and in the dense version every oscillator talks to every other one. That sounds like a strength, but it means the whole system can fall into catastrophic synchronization: everything locks into step, gradients go flat, and learning stalls. Sparser connectivity leaves room for coherent and incoherent activity to coexist. The dynamics stay alive, and the model keeps learning. Connectivity turns out to be a control knob, not a dial you turn up until it stops. Learn more here:

Unconventional AI

34,149 просмотров • 5 дней назад