refresh
2024-03-20T09:59:59.325Z
Thank you for watching ITPro's coverage of the opening keynote at KubeCon 2024 in Paris. Stay tuned for more KubeCon news in the coming days!
2024-03-20T09:58:16.996Z
The audience is thanked for their participation, and just like that, people begin to be excluded from the keynote theater.
(Image credit: Future) 2024-03-20T09:51:44.409Z
And as they leave the stage, it looks like it's time for the closing remarks of the opening keynote speech. Organizers will announce some exciting upcoming events during the conference.
(Image credit: Mirai) 2024-03-20T09:47:22.770Z
Managed model training is the topic of their talk, and they discuss the use of Buildpacks and Kubeflow before moving on to multi-cluster scheduling. Knative and KServe are also platforms used by Bloomberg.
2024-03-20T09:46:46.989Z
Supporting the feature requires a scalable infrastructure and must support the complex machine learning lifecycle. They leverage open source solutions to aid in this process.
2024-03-20T09:46:28.298Z
They talk about the company's flagship product, the Bloomberg terminal. The company processes large amounts of data every day. They talk about how the Bloomberg terminal uses AI (not surprising given the focus of previous keynotes).
2024-03-20T09:41:00.157Z
The next keynote will feature two more speakers. Yuzhui Liu, Bloomberg's team lead, and Leon Chou, his software engineer at Bloomberg.
(Image credit: Mirai) 2024-03-20T09:36:53.218Z
When another speaker bursts into the newsroom, Ibrahim Haddad (via video), the news program is over.
2024-03-20T09:31:20.948Z
Here, Kakodkar speaks more generally about the relationship between cloud native and AI, and the symbiotic relationship between infrastructure and cloud native.
(Image credit: Future) 2024-03-20T09:29:17.149Z
The keynote speech will be switched to mimic a “live report” from outside the conference center.
2024-03-20T09:28:27.505Z
This just… Kubernetes is entering its 20th year. Having learned how to walk and run, he now “jumps” into the AI ​​galaxy. The keynote session will likely feature several different “news broadcasters” or just speakers during the event.
2024-03-20T09:22:40.773Z
Next to the stage is the curiously named Cloud Native News Show (complete with fake news show graphics), led by Nikita Raghunath and Rajas Kakodar.
(Image credit: Future) 2024-03-20T09:20:08.778Z
To conclude, the panel organizers reaffirmed that Kubernetes is becoming the standard for AI platforms, and included a panel discussion on resource allocation, data loading, and the importance of choosing the right dedicated compute for the right model. I also mentioned the department.
2024-03-20T09:14:11.511Z
The question being asked is: What can we do to optimize sustainability? One panelist joined us to tout less usage of computing systems, resulting in lower costs.
2024-03-20T09:12:16.502Z
Computing power seems to be a continuing theme at KubeCon 2024's opening keynote, with panel attention focused heavily on the concepts of GPU sharing and efficient GPU usage.
2024-03-20T09:10:58.429Z
The panel discussion began with questions about how Kubernetes can be simplified, followed by questions about the compute decisions people should make regarding Kubernetes. His GPU-free LLM inference is his one answer to the latter of the two questions.
2024-03-20T09:06:42.833Z
Next to the stage are several people participating in a panel discussion, including Alluxio's Lu Qiu, Google's Susan Wu and Clayton Coleman, Ampere Computing's Victor Jakubiuk, and CERN's Ricardo Rocha.
(Image credit: Future) 2024-03-20T09:06:26.912Z
That's why he announced the company's new KAITO architecture for deploying and running model inference. After a quick check of the products, we're off.
2024-03-20T09:05:32.107Z
Next to the stage will be Jorge Palma, principal PM lead at Microsoft, who will lead a sponsored keynote on building open source platforms for AI. There are many challenges along the way, he says, such as containerizing the model.
2024-03-20T08:54:32.110Z
“Topology-aware placement” is the key to avoiding problems, he says, and understanding fault tolerance and resiliency is important. You need to be able to easily diagnose problems and build in observability.
2024-03-20T08:50:56.543Z
Klues will hand over to Chatterjee to talk about the issues surrounding scaling out GPUs, especially when it comes to training. Meeting the demands of generative AI will require massive amounts of GPUs, he says.
2024-03-20T08:49:13.932Z
They talk about “time slicing,” a technique used to split the demands of AI workloads, and GPU sharing. Kevin gives a little shout-out about other talks at the conference if people want to know more.
2024-03-20T08:48:47.611Z
They talk about how GPUs are at the center of the AI ​​revolution and what it takes to enable GPU support in Kubernetes, support that helps control AI workloads.
2024-03-20T08:38:42.830Z
Two new speakers, NVIDIA Distinguished Engineer Kevin Klues and Engineering Manager Sanjay Chatterjee, took to the stage to talk about GPU-powered AI workloads in Kubernetes.
(Image credit: Mirai) 2024-03-20T08:37:37.876Z
Sharma brings up the term “open source washing” to draw the group's attention to the importance of open source in AI. While speaking about the open source nature of Kubernetes, Morgan also commented that this is influenced by whether the model is open source or not, and LaCroix said it would be helpful if all tools for AI were open source. states.
2024-03-20T08:37:01.054Z
The group will also talk about inference cases, the possibility of using smaller-scale models, and the evolution of the universe in general. They talk about various AI workloads such as training, fine-tuning, and inference.
2024-03-20T08:27:24.991Z
Sharma is focused on bringing together cloud native and AI, and this time he will be joined by Timothée Lacroix, co-founder of Mistral AI, Paiege Bailey, lead product manager at Google, and Jeffery, founder of Ollama. Mr. Morgan is on stage. They discuss the relationship between AI and cloud native.
(Image credit: Mirai) 2024-03-20T08:25:46.920Z
Sharma has developed a visual language model that impressively details a photo of the conference hall itself. Everyone was suitably satisfied.
2024-03-20T08:21:55.576Z
Sharma talks about how cloud native has made the platform more resilient and stable, naming some of the big names associated with cloud native, including Adidas, Disney, and Spotify. The list goes on.
(Image credit: Future)
According to Sharma, this is clearly the biggest KubeCon ever. It's so big that her parents even come here, she jokes.
2024-03-20T08:20:24.979Z
Priyanka Sharma, Executive Director of Cloud Native Computing Foundation (CNCF), appeared on stage to thunderous applause from the audience.
2024-03-20T08:19:57.636Z
2024-03-20T07:59:12.960Z
The keynote speech started with a promotional video!
(Image credit: Mirai) 2024-03-20T07:47:27.936Z
Everyone can now enter the keynote theater, which is quickly filling up.
2024-03-20T07:35:45.240Z
Admission to the keynote theater is not yet permitted, so this is a good time to get acquainted with the conference space.
(Image credit: Future) 2024-03-20T07:30:58.947Z
The weather is getting better! The KubeCon keynote will be held on his third floor of the center.
(Image credit: Mirai) 2024-03-20T07:22:59.474Z
Unfortunately, the Paris Expo Center coffee shop is not yet open, but the registration area is filling up.
(Image credit: Future)
It's a crisp but sunny day here in Paris as people head to the Paris Expo Center for the first day of KubeCon 2024.
(Image credit: Mirai) 2024-03-15T14:59:33.706Z
Welcome to ITPro's live coverage of KubeCon 2024 at the Paris Expo Conference Center. We are on site in Paris, waiting for the opening keynote speech to begin.
In today's session, you'll hear from the Cloud Native Computing Foundation (CNCF) about the latest developments in the cloud native space, as well as community updates and live presentations from many industry participants.
