OpenAI 개발자 데이 런던: 하이라이트와 혁신 공개

OpenAI의 01은 개발자 데이 런던에서 애플리케이션 구축과 실시간 동적 상호작용 등 인상적인 기능을 선보였습니다.

### OpenAI 개발자 데이 런던: 주요 특징 및 혁신 01
* 🚀 런던에서 열린 OpenAI 개발자 데이에서 **01은 인상적인 라이브 데모**를 통해 01의 기능을 선보이며 눈길을 끌었습니다.
* 🤖 이 팀은 **01이 드론을 제어하여 비디오 스트림을 성공적으로 시작**하고 명령을 원활하게 실행할 수 있는 풀스택 애플리케이션을 선보였습니다.
* 📧 특히 다양한 세대를 시각화하는 방식에 있어 **사회적 고정관념을 반영하는 AI 모델에 대한 연구**가 AI Port와 튜링 포스트의 후원으로 주목받았습니다.
* 🌍 또 다른 데모 01에서는 **음성을 사용하여 호텔 예약 및 음식 주문 등 사용자의 여행 계획**을 지원하는 실시간 API 기능을 선보였습니다.
* 📞 음성 상호작용과 실시간 API의 통합으로 **전화로 파이를 대량 주문하는 것과 같은 복잡한 작업**을 수행할 수 있어 일상 생활에서 AI의 실제 적용 사례를 보여줬습니다.
* 📊 이 프레젠테이션에서는 행성에 대한 인터랙티브 학습 코너를 통해 **01이 어떻게 역동적인 실시간 교육 콘텐츠**를 제공할 수 있는지를 강조했습니다.
* 🎤 추가 발표에는 **음성 대화 환경을 위한 새로운 표현력이 풍부한 음성**과 **텍스트 및 오디오 입력 비용의 대폭 할인**이 포함되어 개발자의 접근성을 향상시켰습니다.
* 🔜 OpenAI는 곧 **01 정식 버전**을 출시할 계획이며, 개발자 데이의 청중은 **01 프리뷰에 독점적으로 액세스**하여 AI 개발의 중요한 한 걸음을 내디뎠습니다.

open AI Dev day in London just happened, and 01 was the star of the show. We have some incredible demos to go over today. Let's get right into it. First, the open AI team is going to show off 01 building a full-stack application to control a drone. It is so cool. Check this out. I'll it down right there, and I will go back to the terminal to start the server we just created. Now let's take a look at the running good news. F the battery l, so that's a good sign. I'm trying to start the video stream to see that was up. Great, so now the real test. Take off and see what happens. All right, that worked. So let's see if the buttons work. The events were wired up correctly. I guess the good sign later. Awesome. And also I had surprising but so let's see what deci with this absolutely incredible. Let's go on to the next one. Thanks to the sponsor of this video, AI port and Turing posts. These are great newsletters that cover everything in the world of AI. They break

down the latest AI news and breakthroughs in a really easy-to-read form. They just did a research project where they showed how our own stereotypes get reflected back to us with AI models. For example, they prompted a bunch of text-to-image generators: stable diffusion, mid-journey, Yandex art, and Ernie VG to visualize how different generations, so Gen X, Millennials, Gen Z, etc., might look, and the results were super interesting. Mid-journey seems to think baby boomers are kind of sad and nostalgic; other models show them living their best lives, so really, it just depends on which model you're using, and then that bias is reflected back to you, which if you watch this channel at all, you already know. A lot of models seem to struggle with Generation X, which is kind of weird, and they couldn't quite get a definition of what Generation X actually looks like. But this all just raises a ton of questions about how we're actually training these models, and these are the types of research reports that AI port.

and Turing Post cover. So check out these newsletters. I'll drop links to them in the description below. And thank you again to AI port and Turing post for sponsoring this video. Now back to the video. Next is an absolutely mind-blowing demo 01 using voice helps the person through the platform Wanderlust. Basically find travel plans, look for hotels, and then finally actually finds a bakery, calls the bakery, and puts in an order. It is mind-blowing. Let's watch voice. And so if I type in right here, let's say Lond. Here's London for you. So now, as you can tell, I added a few lines of code with the real-time API clients. So now I can have a conversation with my. So let's take a look. I'm planning my trip to London and Singapore. Let's say Singapore next week. Actually, Singapore is ready for your visit. Awesome. Uh, can you show me some places I could check out while I'm there? Here are some great spots in Singapore. Nice. Uh, I'm thinking I might need a hotel to stay around there. Can

you help me find some hotels in Singapore, nice. Can you show me more details about these three? Here are more details on the hotels. Great. Um, so two things to highlight here on the left as I'm having this conversation with Yak. You can tell the transcript is happening with all of the conversation, but on the right side, the coolest thing is that the real-time API has access to tools. So all of you can use it and call your own functions, your own APIs, your own tools. And in this case, I was able to kind of place pins on the map, annotate the map, zoom on it, fetch details from the internet, and even put some details from this, uh, on the screen. And so we really think you're going to love the real-time API. Your users are going to be delighted to have this new way to interact, whether they're driving, cooking, learning, maybe on the go, trying to be more productive. Um, and now let's try something a bit harder, though. So what if we were able to take this real-time API and the capability that it

has to connect to APIs tools, but let's go outside of just this voice interface. So, uh, what I'm going to do is I'm going to go back into the fun conversation and have another scenario. Let's go back to London now, please, and specifically around the brewery on Chiswell Street. We're back to the brewery on Chiswell Street. Sweet, uh, I have hundreds of developers, uh, here in the room. They might enjoy the little snack. Can you show me some, uh, pie shops around here? Maybe here is in P shops NE. Pretty cool, uh. Can you show me more details about these three things here? Are more details on the P shops? Great, so I'll pause right here. And, uh, now you might have noticed that the second one here is actually demo spot. It's not a real place. The other two are, uh, but what we're going to do now is that because this, uh, second one seems to accept AI orders, what if we were able to use in the context of this wonderless app not just the real-time API but also

toio phone calling capabilities to connect and place phone CES. So for that, I'm going to invite Elan from our developer experience to join me on stage. Thanks. I'm going to jump back into the conversation, and I'll ask one of the last to do some action for me that would leverage this capability of using real-time plus together. So let's check this out. Right, so I think Milan seemed very interesting. Could you please call them for me, and maybe you could see if they could get us, say, like 200 bu, maybe a mix of meat and veggie options, but please keep the budget under F. I'm calling Elan POS now to arrange your order. Hello. Hi, I'm calling onal of re. Yeah, this is ans. How can I help you? Great. I wanted to ask about the kinds of TI you. What do you have? Yeah, we have meat pies, we have veggie pies, we have apple pies, we have all kinds of pies. Perfect. How much would 100 weat

andaran that that's that's a lot of that'll come out to around I don't know 12200 lb. That's overall budget. I don't know if I go 800 but I'll meet you at 1,000 if you guys feature us at your event. All right, okay so 1,000 can uh yes he said 100 meat 100 vegetarian. I' for 100 weat and 100 vegetarian P. When can we expect the differ oh um where where are you located to the [Music] Brewer. Oh for sure we're actually right next door. Fantastic thank you. We look it'll be there like 43 seconds goodbye. Next you're going to see the power of the realtime API that is essentially advanced voice mode and the interaction between a human and whatever the human wants to learn and in this case planets and some information about the planets. Let's watch hey so can you tell me what is the biggest planet in the solar system. Absolutely the biggest planet in

our solar system is Jupiter. It's so massive it has more than twice the mass of all the other planets combined. Jupiter is the giant among giants. Okay, great, so real-time interaction works, and I think it's going to be a more fun way of learning, but let's see if we can make it a bit more interactive. Hey, I'm curious about Earth. Sure thing. Earth is the third planet from the Sun and the only one known to support life. It's packed with diverse ecosystems, abundant water, and a protective atmosphere that shields us from harmful solar radiation. Do you want to know more about Earth's characteristics? Well, I see that there's actually a lot of water there, so can you tell me how much land is there versus water? Absolutely. Approximately 71% of Earth's surface is covered by water, while the remaining 29% is land. Our Blue Marble truly lives up to its name. Let me show you this on a pie chart. Here's the breakdown of Earth's surface. Okay, great, thank you. I actually have a question about

the Mars now Mars the red planet. What would you like to know? So I heard there was a lot of volcano there. Can you tell me which one is the biggest? Absolutely. The biggest volcano on Mars and in the entire solar system is Olympus Mons. It's about 13.6 miles, nearly three times the height of Mount Everest. Would you like more information on Olympus Mons? Yeah, can you tell me like what are the three biggest volcanoes and show me their height? Certainly. The three biggest volcanoes on Mars are: one, Olympus Mons, 13.6 miles, 22 km high; two, Arsia Mons, about 11.3 miles, 18.1 km high; three, Ascraeus Mons, about 11.2 miles, 18 km high. Okay, great, thank you. How incredible was that? Keep in mind all of that is dynamically generated. That is not predefined anywhere, so it really shows the vision of what the future of education can be. Plus, OpenAI had a couple of other smaller announcements. Let's go over those now. You can now build speech-to-speech experiences with five

new voices, which are much more expressive and steerable, and they're lowering the price by using prompt caching. Cash text inputs are discounted 50%, and cashed audio inputs are discounted 80%. Huge discounts towards intelligence too cheap to meter, and everybody in the audience is getting access to 01 preview. Some thought it was the full 01 model, but Olivier Godman, who is the head of product for OpenAI, says, We're giving audience access to 01 preview. We're still working on the full 01 and plan to release it soon. And Jimmy Apple's, of course, right there to comment. Now it's a little weird because we've already had access to 01 preview, so maybe Europeans haven't had access to 01 preview yet. Maybe that's what's going on, but either way, very cool. So those are the major updates from Dev Day London. If you enjoyed this video, please consider giving a like and subscribe, and I'll see you in the next one.

o1 Insane LIVE Demos - o1 FULL Coming Soon! (OpenAI Dev Day London)

Artificial Intelligence (AI), Open Source, Generative Art, AI Art, Futurism, ChatGPT, Large Language Models (LLM), Machine Learning, Technology, Coding, Tutorials, AI News, and more

** Exclusive Pinecone Partner **

Matthew Berman

런던 OpenAI 개발자 데이의 하이라이트

OpenAI의 01은 드론을 제어하는 풀스택 애플리케이션을 인상적인 데모와 함께 선보였습니다.

드론이 이륙하는 동안 비디오 스트림과 버튼 컨트롤이 원활하게 작동했습니다.

이 행사의 후원사인 AI Port와 튜링 포스트는 고정관념을 반영한 AI 모델에 대한 연구 프로젝트를 발표했으며, 다양한 세대를 위한 흥미로운 결과를 보여주었습니다.

음성 인식 여행 계획의 놀라운 데모

데모에서는 사용자가 여행을 계획하고, 호텔을 찾고, 베이커리에서 주문할 수 있도록 도와주는 음성 인식 플랫폼인 Wanderlust를 선보입니다.

실시간 API 클라이언트는 지도 주석 및 인터넷 세부 정보 가져오기와 같은 도구를 제공하여 원활한 대화와 상호 작용을 가능하게 합니다.

이 데모는 이동 중에도 새롭고 생산적인 방식으로 상호 작용할 수 있는 실시간 API의 잠재력을 강조합니다.

실시간 주문을 위한 API 연결 및 전화 통화

화자는 런던의 치스웰 스트리트에 있는 양조장에서 근처 파이 가게를 찾고 있습니다.

화자는 파이 가게 중 한 곳에서 AI 주문을 받는 것을 발견하고 앱의 실시간 API와 전화 통화 기능을 사용하여 주문할 계획입니다.

화자의 동료인 Elan은 전화를 걸어 예산을 염두에 두고 고기와 채소 옵션을 혼합한 파이 200개를 주문하도록 초대받습니다.

이 대화에는 육류 100인분과 채식 100인분에 대해 1,000달러를 약정하는 이벤트의 예산 논의가 포함되어 있습니다.

실시간 API는 대화형 학습을 시연하여 가장 큰 행성인 목성과 생명체를 지원하는 것으로 알려진 유일한 행성인 지구에 대한 정보를 제공합니다.

또한 상호작용을 통해 지구 표면의 71%가 물로 덮여 있다는 사실을 시각적 표현을 위해 원형 차트로 보여줍니다.

올림푸스 몬스는 화성은 물론 태양계 전체에서 가장 큰 화산으로, 높이가 13.

1km)의 아르시아 몬스이며, 아스크라에우스 몬스가 11.

2마일(18km)로 그 뒤를 잇고 있습니다.

OpenAI는 또한 새로운 표현력이 풍부한 음성 인식 음성, 할인된 텍스트 및 오디오 입력, 01 프리뷰에 대한 청중 액세스를 발표했습니다.