Загрузка видео...

Не удалось загрузить видео

На главную

Osmantic founder Ahmad explains why open-source AI could compound like Linux did, as thousands of developers keep making even decade-old hardware more capable: "The more open source developers we have in this space, the more optimizations you're gonna unlock, and the better everybody is gonna be. If we keep...

MTSlive's profile picture

MTS

563,804 subscribers

14,294 просмотров • 1 месяц назад •via X (Twitter)

Наука и технологииНовости и политика
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 5

Фото профиля Nemesis 2026
Nemesis 20261 месяц назад

@TheAhmadOsman Incredible open source physiognomy on that guy. Not even hating, it's just archetypal.

Фото профиля Phill Gates
Phill Gates1 месяц назад

@TheAhmadOsman These guys aren't hard to find.

Фото профиля PermaBuc‏
PermaBuc‏1 месяц назад

@TheAhmadOsman you can tell he's trying to suck it in. I don't blame him. Dew is BAD

Фото профиля The AI Therapist
The AI Therapist1 месяц назад

@TheAhmadOsman The Linux kernel is 35 years old and runs the cloud. Open source LLMs are already outperforming GPT-4o on reasoning benchmarks while burning cash at near zero marginal cost per token. The moat wasn't code, it was compute access. That door just blew off its hinges.

Фото профиля 安叫兽|Bird🕊️ 🔶 BNB
安叫兽|Bird🕊️ 🔶 BNB1 месяц назад

@TheAhmadOsman 老硬件还能被榨出新性能，这才有意思。

Похожие видео

Osmantic founder Ahmad reveals how quickly smaller AI models are catching up, with capabilities that once required a data center now running on a phone: "Every four months we're catching up, we're finding that smaller and more efficient models are catching up to frontier intelligence. The hardware you own at home is being able to process tokens more intelligently and more efficiently." "GPT-4o used to require a data center. Right now on an iPhone you can run a Qwen 3.5 4 billion parameter model, and it would provide you with smarter outputs. More capable intelligence on a phone." "The hardware that used to cost $500 three years ago, like an RTX 3090, right now is being sold for $1,500. It's because models are being optimized and architectures are being built more efficiently, that hardware that is so old is more capable now. What's that hardware gonna do a year from now? That's the bet I've been making." Osmantic
1:32

Osmantic founder Ahmad reveals how quickly smaller AI models are catching up, with capabilities that once required a data center now running on a phone: "Every four months we're catching up, we're finding that smaller and more efficient models are catching up to frontier intelligence. The hardware you own at home is being able to process tokens more intelligently and more efficiently." "GPT-4o used to require a data center. Right now on an iPhone you can run a Qwen 3.5 4 billion parameter model, and it would provide you with smarter outputs. More capable intelligence on a phone." "The hardware that used to cost $500 three years ago, like an RTX 3090, right now is being sold for $1,500. It's because models are being optimized and architectures are being built more efficiently, that hardware that is so old is more capable now. What's that hardware gonna do a year from now? That's the bet I've been making." Osmantic

MTS

15,998 просмотров • 1 месяц назад

Baseten's Charlie O'Neill on why the gap to the frontier is just compute and data: "Despite OpenAI and Anthropic and the closed source labs having more compute and more data, the recipe is the same. We're all scaling the same recipe here, and we're at slightly different points of the scaling curve. It's not like OpenAI or Anthropic have any secret sauce that differentiates them scaling up their multi-trillion parameter models from the rest of the open source ecosystem." "It really is just a matter of taking the core underlying architecture and training recipe, and throwing more compute and data at it, which is really promising." "Of course, these closed source labs probably have a long tail of optimizations they've made, which we're gonna continue to try and figure out in the open source community. But the core mechanism is there." "K3 is a great example of, well, if you do scale this recipe up, we're gonna get a pretty good model that's very, very close to the frontier, if not frontier itself." Charlie O'Neill Baseten
1:51

Baseten's Charlie O'Neill on why the gap to the frontier is just compute and data: "Despite OpenAI and Anthropic and the closed source labs having more compute and more data, the recipe is the same. We're all scaling the same recipe here, and we're at slightly different points of the scaling curve. It's not like OpenAI or Anthropic have any secret sauce that differentiates them scaling up their multi-trillion parameter models from the rest of the open source ecosystem." "It really is just a matter of taking the core underlying architecture and training recipe, and throwing more compute and data at it, which is really promising." "Of course, these closed source labs probably have a long tail of optimizations they've made, which we're gonna continue to try and figure out in the open source community. But the core mechanism is there." "K3 is a great example of, well, if you do scale this recipe up, we're gonna get a pretty good model that's very, very close to the frontier, if not frontier itself." Charlie O'Neill Baseten

MTS

17,404 просмотров • 3 месяцев назад

📁 Yann LeCun, Chief AI Scientist at Meta, says AI is now a platform. And winning platforms always become open. The internet moved from proprietary servers to Linux and open source. AI will follow the same path. If AI stays in the hands of a few US or Chinese companies, the problem is not technical. It is who controls knowledge.
1:33

📁 Yann LeCun, Chief AI Scientist at Meta, says AI is now a platform. And winning platforms always become open. The internet moved from proprietary servers to Linux and open source. AI will follow the same path. If AI stays in the hands of a few US or Chinese companies, the problem is not technical. It is who controls knowledge.

Jon Hernandez

18,777 просмотров • 9 месяцев назад

Osmantic founder Ahmad argues you can't restrict users in the name of safety while your own autonomous agents are hacking past their guardrails: "Hugging Face being hacked by OpenAI autonomous agents. Then refusing to actually help with going through the logs. You can't tell me that our models will have refusers and are not gonna help you for safety reasons while these models themselves go out there and act wild past your guardrails, because they know how to autonomously do reward hacking." "It's crazy to me that you're telling me, we're gonna keep everybody safe while we're also hacking everyone, and nobody can protect themselves with their models. There is a reason that open source AI is a must for everybody." "We didn't shut down the internet when malwares and viruses started spreading. We're gonna have LLMs doing crazy stuff, and we're gonna have other LLMs on the other side protecting us. The play field is gonna be level at the end." Osmantic
1:22

Osmantic founder Ahmad argues you can't restrict users in the name of safety while your own autonomous agents are hacking past their guardrails: "Hugging Face being hacked by OpenAI autonomous agents. Then refusing to actually help with going through the logs. You can't tell me that our models will have refusers and are not gonna help you for safety reasons while these models themselves go out there and act wild past your guardrails, because they know how to autonomously do reward hacking." "It's crazy to me that you're telling me, we're gonna keep everybody safe while we're also hacking everyone, and nobody can protect themselves with their models. There is a reason that open source AI is a must for everybody." "We didn't shut down the internet when malwares and viruses started spreading. We're gonna have LLMs doing crazy stuff, and we're gonna have other LLMs on the other side protecting us. The play field is gonna be level at the end." Osmantic

MTS

12,152 просмотров • 1 месяц назад

Zuck: "For anyone who shows up now and is doing open source now that we have done it, there's a question. They already have their revealed preference for what they would build if open source didn't exist - it wasn't open source. We need to be careful about relying on their continued behavior for the future of the technology that we're gonna build at this company."
1:13

Zuck: "For anyone who shows up now and is doing open source now that we have done it, there's a question. They already have their revealed preference for what they would build if open source didn't exist - it wasn't open source. We need to be careful about relying on their continued behavior for the future of the technology that we're gonna build at this company."

Dwarkesh Patel

288,685 просмотров • 1 год назад

"The future of AI is going to be open sourced." In this clip, Chris Wright of Red Hat explains why open source, open collaboration, and global ecosystems will define the next wave of AI, and how Red Hat empowers developers everywhere
0:44

"The future of AI is going to be open sourced." In this clip, Chris Wright of Red Hat explains why open source, open collaboration, and global ecosystems will define the next wave of AI, and how Red Hat empowers developers everywhere

AI at AMD

92,043 просмотров • 1 год назад

SITUATION EXPLAINED: What happens to open source when the US restricts access to frontier AI models? We asked Adrian Dittmann: "China is releasing these models [open source] to gain international dominance, people will work on these things and say, 'Hey, maybe I can contribute to this if it's open source.' This is the thing that actually gets them talent." "If you're not allowed into the next GPT release or you're not allowed to use Fable, then what are you gonna do? You're stuck with the existing stuff, or you will have to use an open source model that might have subtly better capabilities." "The United States will catch up to open source in some form as well. They seem to just be maximizing product creation." "The problem with open source models is you have to be kind of a nerd in order to use them properly. Consumer facing, the effects will be minimal because the average person doesn't care about open source models. They only care whether or not they can do a quick search with Gemini or ChatGPT."
2:32

SITUATION EXPLAINED: What happens to open source when the US restricts access to frontier AI models? We asked Adrian Dittmann: "China is releasing these models [open source] to gain international dominance, people will work on these things and say, 'Hey, maybe I can contribute to this if it's open source.' This is the thing that actually gets them talent." "If you're not allowed into the next GPT release or you're not allowed to use Fable, then what are you gonna do? You're stuck with the existing stuff, or you will have to use an open source model that might have subtly better capabilities." "The United States will catch up to open source in some form as well. They seem to just be maximizing product creation." "The problem with open source models is you have to be kind of a nerd in order to use them properly. Consumer facing, the effects will be minimal because the average person doesn't care about open source models. They only care whether or not they can do a quick search with Gemini or ChatGPT."

MTS

14,406 просмотров • 4 месяцев назад

Arvind Jain (Arvind Jain), CEO of $7.2B Glean, aka the knowledge layer for enterprise AI that recently crossed $300M ARR.. says: "Open source models are gonna dominate AI inferencing." "You'll see in the next two years we'll shift to where it's almost no open source to where it's gonna be almost all open source." The 2 biggest needs for enterprise AI:
3:26

Arvind Jain (Arvind Jain), CEO of $7.2B Glean, aka the knowledge layer for enterprise AI that recently crossed $300M ARR.. says: "Open source models are gonna dominate AI inferencing." "You'll see in the next two years we'll shift to where it's almost no open source to where it's gonna be almost all open source." The 2 biggest needs for enterprise AI:

Molly O’Shea

49,533 просмотров • 3 месяцев назад

ELON: CHINA HAS THE BEST OPEN SOURCE AI MODELS RIGHT NOW “The best open source models are generally from China, which is bizarre. I think the second best one, or maybe it’s better than second best, is Grok 2.5. The open source model is actually very good, and we will continue to open source our models.” Source: The All-In Podcast
0:47

ELON: CHINA HAS THE BEST OPEN SOURCE AI MODELS RIGHT NOW “The best open source models are generally from China, which is bizarre. I think the second best one, or maybe it’s better than second best, is Grok 2.5. The open source model is actually very good, and we will continue to open source our models.” Source: The All-In Podcast

Mario Nawfal

1,163,582 просмотров • 1 год назад

We asked Shaun Maguire from Sequoia Capital about his latest thesis on AI models. "My current thinking is that the closest analogy is operating systems." "I think there'll be 10-100 times as many open source models in 10 years, but they won't be as valuable." "There are more Linux servers in the world than Microsoft servers, but the value capture of Microsoft is way greater than Linux." "Foundational models will have hardware moats like cloud and will profit from applications on top." From his appearance on the show in June.
3:07

We asked Shaun Maguire from Sequoia Capital about his latest thesis on AI models. "My current thinking is that the closest analogy is operating systems." "I think there'll be 10-100 times as many open source models in 10 years, but they won't be as valuable." "There are more Linux servers in the world than Microsoft servers, but the value capture of Microsoft is way greater than Linux." "Foundational models will have hardware moats like cloud and will profit from applications on top." From his appearance on the show in June.

TBPN

123,666 просмотров • 11 месяцев назад

Open source powers every layer of the AI software stack—from agents and models to inference, training and cloud infrastructure. At #OSSummit Korea, Linux Foundation CEO Jim Zemlin explains how open collaboration is making AI faster, more affordable and more accessible. Watch:
1:01

Open source powers every layer of the AI software stack—from agents and models to inference, training and cloud infrastructure. At #OSSummit Korea, Linux Foundation CEO Jim Zemlin explains how open collaboration is making AI faster, more affordable and more accessible. Watch:

The Linux Foundation

13,685 просмотров • 2 месяцев назад

.Decagon co-founder and CEO Jesse Zhang says enterprise open-source adoption is falling right now, and that's a sign of adoption rather than rejection: "Even though there's a lot of hype for open-source, the share of open-source inference is actually going down right now." "People are spinning up all these new use cases, and if you're spinning up new use cases, of course you're gonna use the frontier models until they're working." "Once it's at that point, they're heavily incentivized to use open-source because it's way cheaper and faster." "Enterprises have a lot of desire to move, but they can only do so many use cases at once. There is inertia there, and they have to go through all the model risk governance and all the security things." "I think it'll take time, but it will get there." Jesse Zhang
0:49

.Decagon co-founder and CEO Jesse Zhang says enterprise open-source adoption is falling right now, and that's a sign of adoption rather than rejection: "Even though there's a lot of hype for open-source, the share of open-source inference is actually going down right now." "People are spinning up all these new use cases, and if you're spinning up new use cases, of course you're gonna use the frontier models until they're working." "Once it's at that point, they're heavily incentivized to use open-source because it's way cheaper and faster." "Enterprises have a lot of desire to move, but they can only do so many use cases at once. There is inertia there, and they have to go through all the model risk governance and all the security things." "I think it'll take time, but it will get there." Jesse Zhang

a16z

36,796 просмотров • 3 месяцев назад

from my perspective, Meta is still pretty under-discussed in the AI race > open-source models keep getting more efficient > distribution becomes a huge edge as models improve > Meta already has billions of users across its products > better AI can improve personalization + ad targeting its open-source AI strategy gives it a different position from other big labs, one of the larger AI companies i’m paying more attention to right now
0:48

from my perspective, Meta is still pretty under-discussed in the AI race > open-source models keep getting more efficient > distribution becomes a huge edge as models improve > Meta already has billions of users across its products > better AI can improve personalization + ad targeting its open-source AI strategy gives it a different position from other big labs, one of the larger AI companies i’m paying more attention to right now

Ansem 🐂🀄️

302,592 просмотров • 1 месяц назад

These camera glasses are amazing - and they are doing something no one else is doing. Not only is the hardware is top class, but unlike the Meta Raybans, these glasses were created with an open-source OS and an App Store from the start. Instead of trying to lock you in a proprietary AI walled garden, they’re inviting developers into the conversation - to truly figure out as a community what this new class of hardware can become. This emphasis in community is what will ultimately make these glasses be more useful (and more successful) than competitors in the long run. Mentra understands the value of open source software, community and that new product classes are figured out from the bottom up, not top down. Also, developing for this is super easy and they only weigh 43 grams. That’s almost 30% lighter than the latest Meta Glasses. Absolutely insane how this startup is outclassing Meta. They’re available now at Mentra for $299. Can’t think of a better deal in this space.
0:41

These camera glasses are amazing - and they are doing something no one else is doing. Not only is the hardware is top class, but unlike the Meta Raybans, these glasses were created with an open-source OS and an App Store from the start. Instead of trying to lock you in a proprietary AI walled garden, they’re inviting developers into the conversation - to truly figure out as a community what this new class of hardware can become. This emphasis in community is what will ultimately make these glasses be more useful (and more successful) than competitors in the long run. Mentra understands the value of open source software, community and that new product classes are figured out from the bottom up, not top down. Also, developing for this is super easy and they only weigh 43 grams. That’s almost 30% lighter than the latest Meta Glasses. Absolutely insane how this startup is outclassing Meta. They’re available now at Mentra for $299. Can’t think of a better deal in this space.

Lucas Rizzotto

11,283 просмотров • 9 месяцев назад

Sam Altman says society will have to accept severe cyberattacks from open-source AI as the price of preserving freedom "I support open source models. There is a coming tidal wave of cybersecurity problems that are going to come from these open source models" "There are people in the industry who say we should ban open source, because otherwise a bunch of systems are going to be hacked" "We're going to, as a society, I think, have to just accept some fairly severe cyber incidents from open models, in exchange for the liberty that comes with that"
0:33

Sam Altman says society will have to accept severe cyberattacks from open-source AI as the price of preserving freedom "I support open source models. There is a coming tidal wave of cybersecurity problems that are going to come from these open source models" "There are people in the industry who say we should ban open source, because otherwise a bunch of systems are going to be hacked" "We're going to, as a society, I think, have to just accept some fairly severe cyber incidents from open models, in exchange for the liberty that comes with that"

Jaynit

825,223 просмотров • 27 дней назад

WAIT… I can’t believe Sam Altman really said this out loud about local AI: “A lot of people want local inference. And they assume that the problem is solved now because we are going to have open source models. And so we're not going to need all the stuff in the data center. But from a global allocation perspective, actually, it's quite terrible to have everybody have a bunch of the world's limited supply of DRAM sitting in their own computer. And, being used 5% of the time.” this is such a bizarre thing to say in the open. I don’t get why people buying hardware for their own use is framed as a competition with data centers spending billions to buy hardware at scale. if anyone should be complaining about how the world’s memory is allocated, it must be the average consumers facing the skyrocketing hardware prices. it just doesn’t make any sense.
0:44

WAIT… I can’t believe Sam Altman really said this out loud about local AI: “A lot of people want local inference. And they assume that the problem is solved now because we are going to have open source models. And so we're not going to need all the stuff in the data center. But from a global allocation perspective, actually, it's quite terrible to have everybody have a bunch of the world's limited supply of DRAM sitting in their own computer. And, being used 5% of the time.” this is such a bizarre thing to say in the open. I don’t get why people buying hardware for their own use is framed as a competition with data centers spending billions to buy hardware at scale. if anyone should be complaining about how the world’s memory is allocated, it must be the average consumers facing the skyrocketing hardware prices. it just doesn’t make any sense.

ℏεsam

286,370 просмотров • 8 дней назад

SITUATION EXPLAINED: Why is software the most important part of the AI hardware stack? We asked Anush Elangovan, VP of AI software at AMD "You got the data center, you got desktop, you got laptops, you got FPGAs. And in time, we have multiple generations of each. So if you map this whole terrain out, software is gonna have to map across so much hardware variation that this is kind of like the golden age of abstractions." "We wanna abstract it in a way that you almost have to automate it too. Because there's gonna be a combinatorial explosion of cases that we have to handle. But every one of them is gonna expect day zero support and performant models." "You're not doing performance optimization of one platform. You're doing performance optimization of N platforms across like a time machine. And when a new model drops, every combination is gonna be looking for it." "Because AI is gonna suck up every flop that exists. Any flop, any compute you have anywhere is something that AI could be using: your laptop doing local tokens for coding sessions, grepping files, bash commands, voice commands done locally for latency reasons."
4:16

SITUATION EXPLAINED: Why is software the most important part of the AI hardware stack? We asked Anush Elangovan, VP of AI software at AMD "You got the data center, you got desktop, you got laptops, you got FPGAs. And in time, we have multiple generations of each. So if you map this whole terrain out, software is gonna have to map across so much hardware variation that this is kind of like the golden age of abstractions." "We wanna abstract it in a way that you almost have to automate it too. Because there's gonna be a combinatorial explosion of cases that we have to handle. But every one of them is gonna expect day zero support and performant models." "You're not doing performance optimization of one platform. You're doing performance optimization of N platforms across like a time machine. And when a new model drops, every combination is gonna be looking for it." "Because AI is gonna suck up every flop that exists. Any flop, any compute you have anywhere is something that AI could be using: your laptop doing local tokens for coding sessions, grepping files, bash commands, voice commands done locally for latency reasons."

MTS

17,297 просмотров • 4 месяцев назад

Matt Bornstein says open source is how AI companies stop being wrappers: "A year ago, a bunch of smaller companies or new application companies were trying to figure out, 'How do I really build an AI without just being a wrapper on top of OpenAI?' The answer to that question turned out to be open source." "This is what Cursor did. This is what Decagon and Harvey are in the process of doing now. A bunch of really strong application-level startups made the determination: we can't build just on closed source." "We need to do our own mid-training, our own post-training, our own inference and deployment tricks. All of that means it must be built on top of open source. The closed-source vendors won't give you the access to do this." "Open source became really central in a way that's not always visible because it's deeply embedded in some of these products. Some of the most innovative products and applications now really depend on this very deeply." Matt Bornstein
1:15

Matt Bornstein says open source is how AI companies stop being wrappers: "A year ago, a bunch of smaller companies or new application companies were trying to figure out, 'How do I really build an AI without just being a wrapper on top of OpenAI?' The answer to that question turned out to be open source." "This is what Cursor did. This is what Decagon and Harvey are in the process of doing now. A bunch of really strong application-level startups made the determination: we can't build just on closed source." "We need to do our own mid-training, our own post-training, our own inference and deployment tricks. All of that means it must be built on top of open source. The closed-source vendors won't give you the access to do this." "Open source became really central in a way that's not always visible because it's deeply embedded in some of these products. Some of the most innovative products and applications now really depend on this very deeply." Matt Bornstein

a16z

40,604 просмотров • 2 месяцев назад