Загрузка видео...

Не удалось загрузить видео

На главную

I got a diffusion model to run really fast, glued it to a LLM, and it blows my mind how much of the interactivity we expect from software (editing text, dragging things around, clicking on "links") just works™️ in pixel space. The future of software is bright

zan2434's profile picture

Zain Shah

20,511 subscribers

22,649 просмотров • 2 месяцев назад •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

Jensen Huang at GTC: "I expect data center build out to reach a trillion dollars, and I am fairly certain we're going to reach that very soon." "General purpose computing has run its course, and that we need a new computing approach." "In the past we wrote the software and we ran it on computers. In the future, the computer is going to generate the tokens for the software." "The computer has become a generator of tokens, not a retrieval of files."
1:47

Jensen Huang at GTC: "I expect data center build out to reach a trillion dollars, and I am fairly certain we're going to reach that very soon." "General purpose computing has run its course, and that we need a new computing approach." "In the past we wrote the software and we ran it on computers. In the future, the computer is going to generate the tokens for the software." "The computer has become a generator of tokens, not a retrieval of files."

vitrupo

55,974 просмотров • 1 год назад

"In the future, you won't need any software at all. The neural network will completely replace it." - stephen balaban, co-founder of Lambda "Here's the idea: instead of generating a program, just go to ChatGPT and say, 'Hi, please behave like a calculator or a spreadsheet.'" "'Generate an ASCII user interface for me. And I want you to essentially just respond, implement the logic of that program in your mind.'" "This is what I call neural software. Normal software is really brittle. 'Neural software' can't have bugs, only misunderstandings." "Large language models will be the software you interact with." From Stephen's appearance on the show in May.
1:42

"In the future, you won't need any software at all. The neural network will completely replace it." - stephen balaban, co-founder of Lambda "Here's the idea: instead of generating a program, just go to ChatGPT and say, 'Hi, please behave like a calculator or a spreadsheet.'" "'Generate an ASCII user interface for me. And I want you to essentially just respond, implement the logic of that program in your mind.'" "This is what I call neural software. Normal software is really brittle. 'Neural software' can't have bugs, only misunderstandings." "Large language models will be the software you interact with." From Stephen's appearance on the show in May.

TBPN

28,665 просмотров • 10 месяцев назад

Results from a failed attempt to train a latent diffusion model on the results of my gabor image fitting process. The model doesn't do what it was supposed to, but I don't really mind what it's doing.
0:56

Results from a failed attempt to train a latent diffusion model on the results of my gabor image fitting process. The model doesn't do what it was supposed to, but I don't really mind what it's doing.

Andrew 😵‍💫

58,435 просмотров • 4 месяцев назад

Anish Acharya: We're going to see a "YouTube moment for software": "If you think about YouTube 20 years ago—we had lots of video and lots of television, and it was high production quality, and it wasn't clear that we needed more and 20 years later, YouTube's a $550 billion enterprise that would be one of the biggest companies in the world if it was independent." "I think the same thing is going to happen for software. People want to make software, and for the first time they can—and they can distribute it and they can consume it." "Sometimes it's going to be important software. Sometimes it's going to be totally trivial. It's going to be software for a bachelor party weekend, software for a joke, software for a prompt. We have this sort of seriousness about software that we had about video and television 20 years ago." "Now it's like—I just took a video on my phone. It's going to be like—I just made an app on my phone. Same energy." Anish Acharya on BILLIONS with Guillaume Moubeche
1:30

Anish Acharya: We're going to see a "YouTube moment for software": "If you think about YouTube 20 years ago—we had lots of video and lots of television, and it was high production quality, and it wasn't clear that we needed more and 20 years later, YouTube's a $550 billion enterprise that would be one of the biggest companies in the world if it was independent." "I think the same thing is going to happen for software. People want to make software, and for the first time they can—and they can distribute it and they can consume it." "Sometimes it's going to be important software. Sometimes it's going to be totally trivial. It's going to be software for a bachelor party weekend, software for a joke, software for a prompt. We have this sort of seriousness about software that we had about video and television 20 years ago." "Now it's like—I just took a video on my phone. It's going to be like—I just made an app on my phone. Same energy." Anish Acharya on BILLIONS with Guillaume Moubeche

a16z

102,995 просмотров • 6 месяцев назад

Candace Owens PROVES Brigitte Macron is her brother with SCIENCE (& that she's NOT a conspiracy theorist) - By using the World Economic Forum's FACE++ AI software. Candace: "It's really good software. The more that I use it the more familiar I am with how it works. Candace seconds later: "I don't know how this software works. But I've learned to trust it."
2:40

Candace Owens PROVES Brigitte Macron is her brother with SCIENCE (& that she's NOT a conspiracy theorist) - By using the World Economic Forum's FACE++ AI software. Candace: "It's really good software. The more that I use it the more familiar I am with how it works. Candace seconds later: "I don't know how this software works. But I've learned to trust it."

Nathan Livingstone (MilkBarTV)

175,639 просмотров • 1 год назад

I went a little crazy with my first 360-drone captured gaussian splat. Of a public park I always bring my daughter to It really is so amazing how much detail a person could capture in 3D with just hobbyist hardware and software
0:57

I went a little crazy with my first 360-drone captured gaussian splat. Of a public park I always bring my daughter to It really is so amazing how much detail a person could capture in 3D with just hobbyist hardware and software

Brad Lynch

35,155 просмотров • 5 месяцев назад

“AI is changing the value prop of software.” We spoke with Metronome CEO Scott Woody about why AI requires a new business model. “In the cloud era…the value of the software scales as your org scales.” “In AI, the software is now doing work on your behalf…so the core value of the software is not who in my org has access to the data, it's what can this software do for me? When the value shifts from access to output, the business model has to follow.
1:36

“AI is changing the value prop of software.” We spoke with Metronome CEO Scott Woody about why AI requires a new business model. “In the cloud era…the value of the software scales as your org scales.” “In AI, the software is now doing work on your behalf…so the core value of the software is not who in my org has access to the data, it's what can this software do for me? When the value shifts from access to output, the business model has to follow.

a16z

23,455 просмотров • 1 год назад

TurboEdit Instant text-based image editing discuss: We address the challenges of precise image inversion and disentangled image editing in the context of few-step diffusion models. We introduce an encoder based iterative inversion technique. The inversion network is conditioned on the input image and the reconstructed image from the previous step, allowing for correction of the next reconstruction towards the input image. We demonstrate that disentangled controls can be easily achieved in the few-step diffusion model by conditioning on an (automatically generated) detailed text prompt. To manipulate the inverted image, we freeze the noise maps and modify one attribute in the text prompt (either manually or via instruction based editing driven by an LLM), resulting in the generation of a new image similar to the input image with only one attribute changed. It can further control the editing strength and accept instructive text prompt. Our approach facilitates realistic text-guided image edits in real-time, requiring only 8 number of functional evaluations (NFEs) in inversion (one-time cost) and 4 NFEs per edit. Our method is not only fast, but also significantly outperforms state-of-the-art multi-step diffusion editing techniques.
0:22

TurboEdit Instant text-based image editing discuss: We address the challenges of precise image inversion and disentangled image editing in the context of few-step diffusion models. We introduce an encoder based iterative inversion technique. The inversion network is conditioned on the input image and the reconstructed image from the previous step, allowing for correction of the next reconstruction towards the input image. We demonstrate that disentangled controls can be easily achieved in the few-step diffusion model by conditioning on an (automatically generated) detailed text prompt. To manipulate the inverted image, we freeze the noise maps and modify one attribute in the text prompt (either manually or via instruction based editing driven by an LLM), resulting in the generation of a new image similar to the input image with only one attribute changed. It can further control the editing strength and accept instructive text prompt. Our approach facilitates realistic text-guided image edits in real-time, requiring only 8 number of functional evaluations (NFEs) in inversion (one-time cost) and 4 NFEs per edit. Our method is not only fast, but also significantly outperforms state-of-the-art multi-step diffusion editing techniques.

AK

16,062 просмотров • 2 лет назад

Benedict Evans says AI means more software, not less: "There's going to be a bunch of stuff you could do with software that you just couldn't do before at all. And so there will be more competition." "If you think about the enterprise software fleet today, you've got three buckets... big iron horizontal systems, so SAP and Workday and your CRM... vertical software... and then in the middle, this fuzzy improvised space of Excel and email and the shared file system." "Now you add ChatGPT to that. Do you do that in an LLM? Is there an LLM tool that means you can do that in Salesforce where you couldn't do it before, or you can do it in your vertical app that you couldn't do before?" "All of which is a long way of saying, what does this do to software? And it must be more software. Way more software." Benedict Evans with Erik Torenberg
1:36

Benedict Evans says AI means more software, not less: "There's going to be a bunch of stuff you could do with software that you just couldn't do before at all. And so there will be more competition." "If you think about the enterprise software fleet today, you've got three buckets... big iron horizontal systems, so SAP and Workday and your CRM... vertical software... and then in the middle, this fuzzy improvised space of Excel and email and the shared file system." "Now you add ChatGPT to that. Do you do that in an LLM? Is there an LLM tool that means you can do that in Salesforce where you couldn't do it before, or you can do it in your vertical app that you couldn't do before?" "All of which is a long way of saying, what does this do to software? And it must be more software. Way more software." Benedict Evans with Erik Torenberg

a16z

36,975 просмотров • 3 месяцев назад

TokenFlow: Consistent Diffusion Features for Consistent Video Editing paper page: The generative AI revolution has recently expanded to videos. Nevertheless, current state-of-the-art video models are still lagging behind image models in terms of visual quality and user control over the generated content. In this work, we present a framework that harnesses the power of a text-to-image diffusion model for the task of text-driven video editing. Specifically, given a source video and a target text-prompt, our method generates a high-quality video that adheres to the target text, while preserving the spatial layout and motion of the input video. Our method is based on a key observation that consistency in the edited video can be obtained by enforcing consistency in the diffusion feature space. We achieve this by explicitly propagating diffusion features based on inter-frame correspondences, readily available in the model. Thus, our framework does not require any training or fine-tuning, and can work in conjunction with any off-the-shelf text-to-image editing method. We demonstrate state-of-the-art editing results on a variety of real-world videos.
0:36

TokenFlow: Consistent Diffusion Features for Consistent Video Editing paper page: The generative AI revolution has recently expanded to videos. Nevertheless, current state-of-the-art video models are still lagging behind image models in terms of visual quality and user control over the generated content. In this work, we present a framework that harnesses the power of a text-to-image diffusion model for the task of text-driven video editing. Specifically, given a source video and a target text-prompt, our method generates a high-quality video that adheres to the target text, while preserving the spatial layout and motion of the input video. Our method is based on a key observation that consistency in the edited video can be obtained by enforcing consistency in the diffusion feature space. We achieve this by explicitly propagating diffusion features based on inter-frame correspondences, readily available in the model. Thus, our framework does not require any training or fine-tuning, and can work in conjunction with any off-the-shelf text-to-image editing method. We demonstrate state-of-the-art editing results on a variety of real-world videos.

AK

143,469 просмотров • 3 лет назад

Google just released Muse: a text to image model achieving state of the art performance. It is significantly more efficient than diffusion models. Muse also allows for inpainting, outpainting, and mask-free editing without finetuning or inverting the model.
0:33

Google just released Muse: a text to image model achieving state of the art performance. It is significantly more efficient than diffusion models. Muse also allows for inpainting, outpainting, and mask-free editing without finetuning or inverting the model.

bleedingedge.ai

183,750 просмотров • 3 лет назад

.Amjad Masad : Creating software is going to be one of the most exciting areas. The average person will soon be as capable as what a senior software engineer is today, as we shift from doing the work to directing it. Instead of coding line by line, you’ll run multiple agents in parallel, building the frontend, refactoring the database, shipping features, all in the background. On top of that, you’ll have a clean interface to design and interact with AI more creatively. Software creation is moving from execution to orchestration, and soon, it won’t be a specialised skill, it’ll be the baseline.
0:47

.Amjad Masad : Creating software is going to be one of the most exciting areas. The average person will soon be as capable as what a senior software engineer is today, as we shift from doing the work to directing it. Instead of coding line by line, you’ll run multiple agents in parallel, building the frontend, refactoring the database, shipping features, all in the background. On top of that, you’ll have a clean interface to design and interact with AI more creatively. Software creation is moving from execution to orchestration, and soon, it won’t be a specialised skill, it’ll be the baseline.

Extraordinary

19,743 просмотров • 6 месяцев назад

Tom Kim speaks on Tiger Woods being a role model 💭 “Obviously playing TGL, being on Tiger’s team and being able to really ask him a lot of questions on certain things and he’s been really helpful a lot of them time. “You know, this was my first win in three years and the first person that text me was Tiger Woods. It just shows you the type of person he is, how much he cares. “Those people that have achieved so much, I’ve been able to listen to their interviews, seeing how they talk and asking them advice. I’ve kind of leaned into the best players in the world. Even on YouTube, just listening to them speaking in their interviews. I’m really trying to pay attention to those things and really trying to blend it in my mind.” #GenesisScottishOpen | #RolexSeries
0:52

Tom Kim speaks on Tiger Woods being a role model 💭 “Obviously playing TGL, being on Tiger’s team and being able to really ask him a lot of questions on certain things and he’s been really helpful a lot of them time. “You know, this was my first win in three years and the first person that text me was Tiger Woods. It just shows you the type of person he is, how much he cares. “Those people that have achieved so much, I’ve been able to listen to their interviews, seeing how they talk and asking them advice. I’ve kind of leaned into the best players in the world. Even on YouTube, just listening to them speaking in their interviews. I’m really trying to pay attention to those things and really trying to blend it in my mind.” #GenesisScottishOpen | #RolexSeries

DP World Tour

30,760 просмотров • 2 месяцев назад

we officially solved software products for AI UGC we already systemized physical products months ago.. skincare, supplements, gadgets, you drop it in and the ad comes out looking perfect software was a different story so it took me a bit of time because the text on screen always came out as gibberish and the app in the video looked nothing like the actual app, so every software ad ended up looking like slop.. but that is no longer a problem, in the V4 system i built a framework specifically for software, this whole video now comes out of 1 prompt pause the video and look at the screen yourself, every single piece of text on the app is readable, which is traditionally extremely hard to do with AI UGC and the exact reason software and app people stayed out of it then watch when she swipes her hand to go to the next screen, the animation is 1:1 with how the app moves natively and you can tell there is no weird AI transition happening so now we have: > max realism > readable and accurate text on screen every time > a 1:1 representation of your actual app or software which means literally ANY product can be done with AI UGC now, ecom, software, apps, info products, all of it there has never been a better time to scale creatives.
0:24

we officially solved software products for AI UGC we already systemized physical products months ago.. skincare, supplements, gadgets, you drop it in and the ad comes out looking perfect software was a different story so it took me a bit of time because the text on screen always came out as gibberish and the app in the video looked nothing like the actual app, so every software ad ended up looking like slop.. but that is no longer a problem, in the V4 system i built a framework specifically for software, this whole video now comes out of 1 prompt pause the video and look at the screen yourself, every single piece of text on the app is readable, which is traditionally extremely hard to do with AI UGC and the exact reason software and app people stayed out of it then watch when she swipes her hand to go to the next screen, the animation is 1:1 with how the app moves natively and you can tell there is no weird AI transition happening so now we have: > max realism > readable and accurate text on screen every time > a 1:1 representation of your actual app or software which means literally ANY product can be done with AI UGC now, ecom, software, apps, info products, all of it there has never been a better time to scale creatives.

Miko

34,517 просмотров • 9 дней назад

Former Goldman Sachs executive Raoul Pal explains how AI is going to eat traditional software/SAAS. If your product is just software, agentic AI can reproduce it on demand, optimize it, and redeploy it to a better market. "Agentic AI means it’s like having Fiverr, a website of experts you can ask any question. It’ll go away and do the task.... Agentic AI will build, design the website, code it, register the domain name, figure out the branding, figure out the marketing, figure out the email list, figure out the whole thing. So then you and I are in competition. You’ve built this incredible new website. I just go to my AI and say, “Love Steven’s website. Can you just build it better. Boom. 3 minutes. How can we be entrepreneurs in software? Now there’s this theory going around that AI is going to eat software, and I kind of get it." ---- From 'The Diary Of A CEO and Raoul Pal The Journey Man' YT channel. (link in comment)
1:11

Former Goldman Sachs executive Raoul Pal explains how AI is going to eat traditional software/SAAS. If your product is just software, agentic AI can reproduce it on demand, optimize it, and redeploy it to a better market. "Agentic AI means it’s like having Fiverr, a website of experts you can ask any question. It’ll go away and do the task.... Agentic AI will build, design the website, code it, register the domain name, figure out the branding, figure out the marketing, figure out the email list, figure out the whole thing. So then you and I are in competition. You’ve built this incredible new website. I just go to my AI and say, “Love Steven’s website. Can you just build it better. Boom. 3 minutes. How can we be entrepreneurs in software? Now there’s this theory going around that AI is going to eat software, and I kind of get it." ---- From 'The Diary Of A CEO and Raoul Pal The Journey Man' YT channel. (link in comment)

Rohan Paul

48,418 просмотров • 5 месяцев назад

Former Goldman Sachs executive Raoul Pal explains how AI is going to eat traditional software/SAAS. If your product is just software, agentic AI can reproduce it on demand, optimize it, and redeploy it to a better market. "Agentic AI means it’s like having Fiverr, a website of experts you can ask any question. It’ll go away and do the task.... Agentic AI will build, design the website, code it, register the domain name, figure out the branding, figure out the marketing, figure out the email list, figure out the whole thing. So then you and I are in competition. You’ve built this incredible new website. I just go to my AI and say, “Love Steven’s website. Can you just build it better. Boom. 3 minutes. How can we be entrepreneurs in software? Now there’s this theory going around that AI is going to eat software, and I kind of get it." ---- From 'The Diary Of A CEO and Raoul Pal The Journey Man' YT channel. (link in comment)
1:11

Former Goldman Sachs executive Raoul Pal explains how AI is going to eat traditional software/SAAS. If your product is just software, agentic AI can reproduce it on demand, optimize it, and redeploy it to a better market. "Agentic AI means it’s like having Fiverr, a website of experts you can ask any question. It’ll go away and do the task.... Agentic AI will build, design the website, code it, register the domain name, figure out the branding, figure out the marketing, figure out the email list, figure out the whole thing. So then you and I are in competition. You’ve built this incredible new website. I just go to my AI and say, “Love Steven’s website. Can you just build it better. Boom. 3 minutes. How can we be entrepreneurs in software? Now there’s this theory going around that AI is going to eat software, and I kind of get it." ---- From 'The Diary Of A CEO and Raoul Pal The Journey Man' YT channel. (link in comment)

Rohan Paul

17,215 просмотров • 2 месяцев назад

People forget Tesla is just as much a hardcore software company as it is a hardcore hardware company. You feel it the second you use the Cybercab. The software is buttery smooth. Everything responds instantly. And bc the Cybercab has cameras surrounding the car, it can constantly see what’s happening around it. The software takes all of that camera data and turns it into a real-time understanding of the world around the car like vehicles, people, roads, obstacles. You can literally watch that digital world being built around it on the screen as you ride. Hardware + software + AI, all built to work together. That vertical integration is Tesla’s superpower. That’s the part people miss.
0:28

People forget Tesla is just as much a hardcore software company as it is a hardcore hardware company. You feel it the second you use the Cybercab. The software is buttery smooth. Everything responds instantly. And bc the Cybercab has cameras surrounding the car, it can constantly see what’s happening around it. The software takes all of that camera data and turns it into a real-time understanding of the world around the car like vehicles, people, roads, obstacles. You can literally watch that digital world being built around it on the screen as you ride. Hardware + software + AI, all built to work together. That vertical integration is Tesla’s superpower. That’s the part people miss.

Teslaconomics

23,937 просмотров • 7 дней назад