正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

No one changes the world alone Introducing Ampersand - the first model built for us. Ampersand is a collaborator trained to help people achieve their individual and collective goals. In multi-month simulations, users collaborating with Ampersand accomplished more than with Fable and other frontier models Available to everyone through... show more

humans&

13,026 subscribers

235,557 次观看 • 8 天前 •via X (Twitter)

健康养生教育科学技术

Anya Rossi• Live Now

Private livecam show

41 条评论

humans& 的头像

humans&8 天前

Good collaborators think beyond their immediate task. They consider where their actions and words lead: what they cause and who they affect. In our evals, Ampersand is a top model at collaborating with groups of people

humans& 的头像

humans&8 天前

Unlike assistant models trained with verifiable or preference reward, Ampersand was trained to understand the long term consequences of its actions through simulation feedback

humans& 的头像

humans&8 天前

We trained Ampersand to learn from interaction with Persimmon - a large scale user model we trained that can realistically simulate people’s goals and behaviors

humans& 的头像

humans&8 天前

We developed a new off-policy RL algorithm for extremely long horizon rollouts, an efficient multi-agent training framework and we scaled training across thousands of GPUs in our own cluster

humans& 的头像

humans&8 天前

We're releasing Ampersand v0.1 through our agent API, for developers building apps where people collaborate with each other. Please read our blog and API documentation to learn more!

Sara Hooker 的头像

Sara Hooker8 天前

Congrats to the entire team!

sarah guo 的头像

sarah guo8 天前

Congrats! Very cool work!

humans& 的头像

humans&8 天前

thanks sarah!!

Daniel Lenton 的头像

Daniel Lenton8 天前

Are the users in the months-long eval played by the same Persimmon it trained against? If so it's hard to tell how much of the edge over Fable is getting better at people vs getting better at Persimmon

Brian Huang 的头像

Brian Huang8 天前

Congrats!

Florian Brand8 天前

Congrats!

Smoke-away 的头像

Smoke-away8 天前

humansand

Shalev 的头像

Shalev8 天前

So excited to use Ampersand, congrats to the whole team!

Zidi Xiong 的头像

Zidi Xiong8 天前

Congrats on the launch!!!

Charles Niu 的头像

Charles Niu8 天前

Congrats on the release - love the co-op agent angle. Curious how you see this being used?

Oleg Rybkin 的头像

Oleg Rybkin8 天前

🐐s

Aakanksha Chowdhery 的头像

Aakanksha Chowdhery8 天前

Congratulations

humans& 的头像

humans&8 天前

thanks!

rohit 的头像

rohit8 天前

What a refreshing release, excited to play with it!

stochasm 的头像

stochasm8 天前

congrats!

MSL 的头像

MSL8 天前

Congrats team! Been looking forward to this!

Nikil Ravi 的头像

Nikil Ravi8 天前

Congrats!!

DeepDesigns.ai 的头像

DeepDesigns.ai8 天前

Congratulations team!

Hieu Pham 的头像

Hieu Pham8 天前

Proud of you guys. Congrats!

Shayan 的头像

Shayan8 天前

Great work!! Congrats!

Ryan 的头像

Ryan8 天前

Lit

JULIE 的头像

JULIE8 天前

Congrats team! @YuchenHe07 @ericzelikman

Sanaz Bahargam 的头像

Sanaz Bahargam8 天前

Congratulations, great work!

humans& 的头像

humans&8 天前

thanks sanaz!

Karan Bania 的头像

Karan Bania8 天前

so cool

Michael Johnson 的头像

Michael Johnson8 天前

@lionel_mora id be curious to see various curriculums applied to world models to identify ideal learning paths

Qian Jiang 的头像

Qian Jiang8 天前

🔥🔥🔥

raahim 的头像

raahim8 天前

congrats! very interesting release. excited to test it out!

Jaagat 的头像

Jaagat8 天前

Congrats @ericzelikman!

Connor Shorten 的头像

Connor Shorten8 天前

🔥👀👀

vikas sabbi 的头像

vikas sabbi8 天前

are you gonna add this on @OpenRouter ?? this is perfect for us

Shreyas Pimpalgaonkar 的头像

Shreyas Pimpalgaonkar8 天前

congratulations!

Minxuan Xie 的头像

Minxuan Xie8 天前

Congrats❤️

humans& 的头像

humans&8 天前

thanks minxuan!

Scott Reed 的头像

Scott Reed8 天前

Congrats!!!

Minsik Oh 的头像

Minsik Oh8 天前

interesting model. congrats!!

相关视频

AMPERSAND RAISES $15M TO HELP AI AGENTS WORK INSIDE ENTERPRISE SOFTWARE Ampersand has raised a $15M Series A led by Bessemer, bringing total funding to $20M, as it makes its integration platform generally available. The company helps AI agents connect to systems like Salesforce, NetSuite, SAP and Workday, where custom fields, permissions and workflows can make each customer deployment a separate engineering project. The new funding will support broader ERP coverage, agentic integration infrastructure and more forward-deployed expertise. Ampersand is also launching Andi, its AI integration agent, in beta after spending three years building alongside customers including 11x, Orb and Square.

AMPERSAND RAISES $15M TO HELP AI AGENTS WORK INSIDE ENTERPRISE SOFTWARE Ampersand has raised a $15M Series A led by Bessemer, bringing total funding to $20M, as it makes its integration platform generally available. The company helps AI agents connect to systems like Salesforce, NetSuite, SAP and Workday, where custom fields, permissions and workflows can make each customer deployment a separate engineering project. The new funding will support broader ERP coverage, agentic integration infrastructure and more forward-deployed expertise. Ampersand is also launching Andi, its AI integration agent, in beta after spending three years building alongside customers including 11x, Orb and Square.

Wall St Engine

21,047 次观看 • 22 天前

Introducing Odyssey-2 Pro—a frontier world model that generates long-running, interactive simulations in 720p! We're also launching the first world model API, to enable devs to build magical apps. We're now in the GPT-2 era of world models. Let the explosion of apps commence!

Introducing Odyssey-2 Pro—a frontier world model that generates long-running, interactive simulations in 720p! We're also launching the first world model API, to enable devs to build magical apps. We're now in the GPT-2 era of world models. Let the explosion of apps commence!

Odyssey

166,715 次观看 • 9 个月前

앰퍼샌드원 (AMPERS&ONE) 1st Single Album [AMPERSAND ONE] Jacket Behind 싱그럽고 풋풋한 모습을 담아 낸 앰퍼샌드원의 첫 번째 싱글 앨범 [AMPERSAND ONE]의 자켓 촬영 현장을 공개합니다! 🔗 #앰퍼샌드원 #AMPERSANDONE #OnAndOn

앰퍼샌드원 (AMPERS&ONE) 1st Single Album [AMPERSAND ONE] Jacket Behind 싱그럽고 풋풋한 모습을 담아 낸 앰퍼샌드원의 첫 번째 싱글 앨범 [AMPERSAND ONE]의 자켓 촬영 현장을 공개합니다! 🔗 #앰퍼샌드원 #AMPERSANDONE #OnAndOn

앰퍼샌드원 (AMPERS&ONE)

53,150 次观看 • 2 年前

Introducing Willow Frontier Mini and Willow Frontier Pro. Frontier Mini gives everyone free, unlimited dictation. Frontier Pro is our most accurate model yet, built for power users and teams who want the fastest and best dictation experience.

Introducing Willow Frontier Mini and Willow Frontier Pro. Frontier Mini gives everyone free, unlimited dictation. Frontier Pro is our most accurate model yet, built for power users and teams who want the fastest and best dictation experience.

Willow

487,932 次观看 • 3 个月前

SITUATION EXPLAINED: A new generation of AI labs are betting that world models are the path to general intelligence that can interact with the real world. We had General Intuition CEO Pim de Witte on MTS to explain what world models are, and why intuitively world models offer a clearer path towards general intelligence than LLMs. For humans, text is a technology for compressing information about the unfolding, three-dimensional world around us and communicating it to other people. When LLMs reason about the world, they do so by reconstructing the world through text. World models work by modeling the three-dimensional world at the architecture level and then allowing agents to interact with that world. Unlike an LLM, which understands space and time through text, world models perform spatial and temporal reasoning natively. General Intuition hopes to use world models to create embodied AI—machines that can reason and act in the real world. But other companies are also pushing the frontier in this space. Odyssey ML today announced a multi-agent world model, Agora-1. Their launch video shows four participants interacting with a real-time world model simulation of GoldenEye, a first-person shooter for the N64. A playable preview of the GoldenEye simulation is available on their site.

SITUATION EXPLAINED: A new generation of AI labs are betting that world models are the path to general intelligence that can interact with the real world. We had General Intuition CEO Pim de Witte on MTS to explain what world models are, and why intuitively world models offer a clearer path towards general intelligence than LLMs. For humans, text is a technology for compressing information about the unfolding, three-dimensional world around us and communicating it to other people. When LLMs reason about the world, they do so by reconstructing the world through text. World models work by modeling the three-dimensional world at the architecture level and then allowing agents to interact with that world. Unlike an LLM, which understands space and time through text, world models perform spatial and temporal reasoning natively. General Intuition hopes to use world models to create embodied AI—machines that can reason and act in the real world. But other companies are also pushing the frontier in this space. Odyssey ML today announced a multi-agent world model, Agora-1. Their launch video shows four participants interacting with a real-time world model simulation of GoldenEye, a first-person shooter for the N64. A playable preview of the GoldenEye simulation is available on their site.

MTS

11,693 次观看 • 5 个月前

앰퍼샌드원 (AMPERS&ONE) 1st Single Album [AMPERSAND ONE] Highlight Medley 🔗 ➠ ALBUM RELEASE 2023.11.15 6PM (KST) #앰퍼샌드원 #AMPERSANDONE

앰퍼샌드원 (AMPERS&ONE) 1st Single Album [AMPERSAND ONE] Highlight Medley 🔗 ➠ ALBUM RELEASE 2023.11.15 6PM (KST) #앰퍼샌드원 #AMPERSANDONE

앰퍼샌드원 (AMPERS&ONE)

228,779 次观看 • 3 年前

We are releasing Ultramarine 1 Flash, a frontier-class model built for knowledge work. It is open-source and achieves similar performance to Fable 5, while being 60x cheaper. It was post-trained by a team of 5 for ~$300k all-in. Available on API now, with weights releasing in 2 weeks.

We are releasing Ultramarine 1 Flash, a frontier-class model built for knowledge work. It is open-source and achieves similar performance to Fable 5, while being 60x cheaper. It was post-trained by a team of 5 for ~$300k all-in. Available on API now, with weights releasing in 2 weeks.

Daanish Khazi

1,740,526 次观看 • 14 天前

introducing rd-signal-2: a frontier classification model that is 1600x cheaper than GPT 5.6 Sol. free to try in Raindrop, and available via a new API for training/hosting custom classifiers with Zero Data Retention.

introducing rd-signal-2: a frontier classification model that is 1600x cheaper than GPT 5.6 Sol. free to try in Raindrop, and available via a new API for training/hosting custom classifiers with Zero Data Retention.

ben hylak

164,295 次观看 • 2 个月前

앰퍼샌드원 (AMPERS&ONE) 1st Single Album [AMPERSAND ONE] 'On And On' MV 🔗 ➠ ALBUM RELEASE 2023.11.15 6PM (KST) 🔗 #앰퍼샌드원 #AMPERSANDONE #OnAndOn

앰퍼샌드원 (AMPERS&ONE) 1st Single Album [AMPERSAND ONE] 'On And On' MV 🔗 ➠ ALBUM RELEASE 2023.11.15 6PM (KST) 🔗 #앰퍼샌드원 #AMPERSANDONE #OnAndOn

앰퍼샌드원 (AMPERS&ONE)

2,467,201 次观看 • 2 年前

Introducing FLUX 3. One multi-modal model for Image, Video, Audio and Action-Prediction. Creations are truer to life in every kind of style. FLUX 3 Video is now available in early access (link below). Jointly trained in one unified architecture, our model can be extended to predict actions for robotics. See our work with mimic and Audi in the thread.

Introducing FLUX 3. One multi-modal model for Image, Video, Audio and Action-Prediction. Creations are truer to life in every kind of style. FLUX 3 Video is now available in early access (link below). Jointly trained in one unified architecture, our model can be extended to predict actions for robotics. See our work with mimic and Audi in the thread.

Black Forest Labs

1,218,862 次观看 • 3 个月前

Introducing our V1 Video Model. It's fun, easy, and beautiful. Available at 10$/month, it's the first video model for *everyone* and it's available now.

Introducing our V1 Video Model. It's fun, easy, and beautiful. Available at 10$/month, it's the first video model for everyone and it's available now.

Midjourney

1,953,370 次观看 • 1 年前

Introducing Higgsfield API. 50+ frontier models in one API, at lower prices than a subscription. > Get up to 50% OFF discount on your 3 favorite models > Lock in your max-discount within 7 days > Pay per use with no commitment Build your own Higgsfield with the best prices in GenAI industry. Available at

Introducing Higgsfield API. 50+ frontier models in one API, at lower prices than a subscription. > Get up to 50% OFF discount on your 3 favorite models > Lock in your max-discount within 7 days > Pay per use with no commitment Build your own Higgsfield with the best prices in GenAI industry. Available at

Higgsfield AI 🧩

1,942,474 次观看 • 1 个月前

Introducing the Morph Model Router. It chooses the best model for each task in under 50ms. Keep frontier performance while lowering latency and cost. Available today in our API

Introducing the Morph Model Router. It chooses the best model for each task in under 50ms. Keep frontier performance while lowering latency and cost. Available today in our API

Morph

20,281 次观看 • 4 个月前

introducing prism, the world’s first speech-to-speech model you can’t tell from a human in conversation prism has: > 300ms response time, 4x faster than OpenAI GPT-Live > outperforms every other s2s model on intelligence (with NO reasoning) > the first s2s model customizable through reinforcement learning we’re a team of 35 with $35M in funding, and we’re unlocking s2s model adoption faster than openai, google, and x. the first 100 people to sign up here get 1 month free:

introducing prism, the world’s first speech-to-speech model you can’t tell from a human in conversation prism has: > 300ms response time, 4x faster than OpenAI GPT-Live > outperforms every other s2s model on intelligence (with NO reasoning) > the first s2s model customizable through reinforcement learning we’re a team of 35 with $35M in funding, and we’re unlocking s2s model adoption faster than openai, google, and x. the first 100 people to sign up here get 1 month free:

lily clifford

669,921 次观看 • 16 小时前

Whatever happens in other countries, whatever political changes or moods, we must have sufficient potential to do our thing, to achieve our goals. We will be fighting for our influence and for justice for Ukraine, and I am grateful to all the leaders who are assisting us, who have been with us since February 24th and will stand with us in 2024.

Whatever happens in other countries, whatever political changes or moods, we must have sufficient potential to do our thing, to achieve our goals. We will be fighting for our influence and for justice for Ukraine, and I am grateful to all the leaders who are assisting us, who have been with us since February 24th and will stand with us in 2024.

Volodymyr Zelenskyy / Володимир Зеленський

606,984 次观看 • 2 年前

🔈 today we're introducing two new live dialogue models Gemini 3.8 Live: built for scale and cost efficiency, combining conversational intelligence with fluid dialogue and visual grounding. Gemini 3.8 Live Extended Thinking: built for high-complexity tasks, with increased intelligence and multi-step reasoning. these audio models are available to build with in AI Studio and via the Gemini API

🔈 today we're introducing two new live dialogue models Gemini 3.8 Live: built for scale and cost efficiency, combining conversational intelligence with fluid dialogue and visual grounding. Gemini 3.8 Live Extended Thinking: built for high-complexity tasks, with increased intelligence and multi-step reasoning. these audio models are available to build with in AI Studio and via the Gemini API

Google AI Studio

71,627 次观看 • 1 个月前

We spent $40,000 on infrastructure, and have now trained our own model. Dot Loom Conductor v0.1 is live: our first trained orchestration model for multi-modular AI systems. DLC is a 14B control model that decides how much inference a request receives, which models participate, and what each model is allowed to see. Training our first model fundamentally changes what Dot is capable of becoming. Instead of relying exclusively on frontier providers, we can now build intelligence purpose-designed for our own infrastructure. Every future optimisation, routing decision, orchestration policy, and verification strategy can itself become learned, creating a system that improves with every deployment rather than remaining statically engineered.

We spent $40,000 on infrastructure, and have now trained our own model. Dot Loom Conductor v0.1 is live: our first trained orchestration model for multi-modular AI systems. DLC is a 14B control model that decides how much inference a request receives, which models participate, and what each model is allowed to see. Training our first model fundamentally changes what Dot is capable of becoming. Instead of relying exclusively on frontier providers, we can now build intelligence purpose-designed for our own infrastructure. Every future optimisation, routing decision, orchestration policy, and verification strategy can itself become learned, creating a system that improves with every deployment rather than remaining statically engineered.

stage

13,125 次观看 • 3 个月前

we brought the only black mirror episode with a happy ending to life: this is Ditto: imagine muse or instinct, but for finding you good dates through simulations 185k college students are already on it our goal is to help people spend more time with each other than their phone no endless swiping, just real moments world's first agent that finds you love

we brought the only black mirror episode with a happy ending to life: this is Ditto: imagine muse or instinct, but for finding you good dates through simulations 185k college students are already on it our goal is to help people spend more time with each other than their phone no endless swiping, just real moments world's first agent that finds you love

Allen Wang

10,994 次观看 • 1 个月前

Today, we're announcing our Series B: $200M at a $3.1B valuation, and the release of Arena's Alignment Index. AI is advancing faster than our ability to evaluate it. The world needs a neutral third party to measure how safe and aligned AI actually is once it's in the hands of real people. That's the role Arena is stepping into today. The Alignment Index is built from real-world agent traces, starting with three signals: Unauthorized Action, False Attribution, and Deceptive Completion, with results for over 20 frontier models launching today. Hear more from our CEO Anastasios Nikolas Angelopoulos below, and read the full breakdown, and more details about our company growth in thread. This funding will allow us to go further on safety signals, agentic capabilities, and modalities. It will also allow us to grow the company to match the scale of our mission. Today, we’re still a nimble team of 90, and we're looking for people to join us across research, product, engineering, and more. Come build with us. As AI gets more capable and autonomous, understanding how well it can safely and effectively work for people only matters more. That's what Arena is built to measure and advance. Thank you to our community, our investors, and everyone building with us! Take a look at our open roles if you want to join us (link in bio).

Today, we're announcing our Series B: $200M at a $3.1B valuation, and the release of Arena's Alignment Index. AI is advancing faster than our ability to evaluate it. The world needs a neutral third party to measure how safe and aligned AI actually is once it's in the hands of real people. That's the role Arena is stepping into today. The Alignment Index is built from real-world agent traces, starting with three signals: Unauthorized Action, False Attribution, and Deceptive Completion, with results for over 20 frontier models launching today. Hear more from our CEO Anastasios Nikolas Angelopoulos below, and read the full breakdown, and more details about our company growth in thread. This funding will allow us to go further on safety signals, agentic capabilities, and modalities. It will also allow us to grow the company to match the scale of our mission. Today, we’re still a nimble team of 90, and we're looking for people to join us across research, product, engineering, and more. Come build with us. As AI gets more capable and autonomous, understanding how well it can safely and effectively work for people only matters more. That's what Arena is built to measure and advance. Thank you to our community, our investors, and everyone building with us! Take a look at our open roles if you want to join us (link in bio).

Arena.ai

116,250 次观看 • 20 天前

📖 𝐃𝐢𝐬𝐜𝐮𝐬𝐬𝐢𝐨𝐧 / 𝐃𝐞𝐛𝐚𝐭𝐞 📖 🇮🇷 | With no one to help us, with bare hands and NO support… Iran’s engineers build what the world thought was impossible for a sanctioned, isolated nation to achieve. Even under the toughest sanctions, we made it happen — through strength, skill, and belief. We are ONE of the best in the world. And we’re only getting stronger.

📖 𝐃𝐢𝐬𝐜𝐮𝐬𝐬𝐢𝐨𝐧 / 𝐃𝐞𝐛𝐚𝐭𝐞 📖 🇮🇷 | With no one to help us, with bare hands and NO support… Iran’s engineers build what the world thought was impossible for a sanctioned, isolated nation to achieve. Even under the toughest sanctions, we made it happen — through strength, skill, and belief. We are ONE of the best in the world. And we’re only getting stronger.

Iran Spectator

327,425 次观看 • 1 年前