What we learned at the world's leading A.I. conference.

What we learned at the world's leading A.I. conference.

NVIDIA's annual A.I. conference for the software development community took place last week, with NVIDIA's CEO Jensen Huang announcing how his company is leading in artificial intelligence, digital twins, robotics, and more.

For our corporate partners, KSV provided an in-depth overview with a breakdown of the top technologies and companies with recommendations for their brands. I invite you to enjoy our highlights from GTC 2024!

NVIDIA coined the term "Accelerated Computing" in 2021, which describes how their GPUs are designed to meet the incredible demands in computational power required to run advanced A.I. models.

Jensen Huang, CEO at NVIDIA

Fun Fact: 30 billion quadrillion calculations (1,000 years with a single PFLOP CPU) is required to train the latest A.I. models! Clearly this is unsustainable, with demands on both the energy grid and environment under increased scrutiny.


The least insightful mention in this post is that A.I. is advancing at an astonishing rate, with Generative A.I. leading the latest wave. This leads us to wonder, what's next? According to panelists at GTC, the next hurdle is to create A.I. agents with deep understanding and reasoning.

Advancements in LLM models in just the last five years:

There's more to it than simply querying a single A.I. model to generate code or an image, however. Companies are embracing multi-agent processes that bring together several task-specific models to work together in concert to achieve a larger objective. One such system was presented by Xavier (Xavi) Amatriain VP of Product, Core ML/AI, Google.

Xavier Amatriain

There are lots of ways that multi-agent designs can be used to provide deeper utility for customer interactions. Think about how one A.I. agent can be an expert at pricing negotiations with a consumer, while another can be used to estimate delivery times by accessing local store and warehouse data.

For marketing workflows, multi-agent designs are even more compelling and straightforward. Perhaps one A.I. agent writes multiple versions of copy to prompt another to generate images, while a third agent reviews the imagery for brand safety and hallucinations before sending approved imagery to the creative team.

Action items for your company:

?? Learn about the latest LLM models and their capabilities - here's a start!

?? Consider which LLM models are best for completing specific tasks.

?? Determine which data sets are needed while being mindful of data privacy.

Chief Technology Officers are rapidly embedding A.I. capabilities into their operations. This is only increasing their desire for employees to be trained on how to work in tandem with their new A.I. tools and coworkers. For employees willing to lean in and learn how to harness A.I. to augment their role and become more effective, there are significant gains in productivity and output to be had.

Training new sales hires in a single day with the assist of "virtual experts" is already a reality using Gen A.I. at our company. - Chris Bedi , Chief Digital Information Officer, ServiceNow
Skills today are more important than a job and title. - Sabry Tozin VP - Engineering, LinkedIn

At L'Oréal, they know Gen A.I. will impact all aspects of their business, so they appointed a central task-force, across HR, R&I, Technology, Digital, & Marketing.

Action items for your company:

?? Appoint a Center of Excellence for Gen A.I. to work across departments.

?? Determine which workflows & skills can be augmented with A.I.

?? Train-up your teams and establish a "test and learn" culture in your company.

Artists around the world are pushing Gen A.I. to new extremes, using the technology to create gorgeous artwork and 3D fantasy worlds. What's striking is not simply that they are using the technology, but that the engineering communities are working in tandem with the artists to push the envelope.

This will create entirely new creative roles where mastering Gen A.I. models will become a mandatory skill set for creative employees at brands and agencies.

Vanessa A Rosa
"The level of conceptual understanding that the engineers have about art is just as deep as someone like me." - Vanessa Rosa, Artist, Little Martians - Mars College

Beyond images and videos, Gen A.I. can be a central piece of technology for creating 3D worlds and objects. Refik Anadol Studios (RAS) used Gen A.I. to generate never-before-seen experiences like the one they debuted below.

Refik Anadol Studios (RAS)

Action items for your company:

?? Encourage your creatives to test and learn new Gen A.I. tools.

?? Request that your agencies incorporate Gen A.I. as part of the creative process.

?? Create a safe space for employees to express themselves and create A.I. art.

Corporations have entered partnerships with tech providers and their agencies to streamline content production workflows with Gen A.I. While companies like L'Oreal are currently engaged in a "test and learn" phase of adoption (according to Chief Digital & Marketing Officer, Asmita Dubey), they are already experiencing the benefits of faster content creation and increased output.

Before rolling out across a portfolio of global brands, A.I. models must be trained on a brand's visual identity to ensure that creative outputs match a brand's style guide while also adopting rapid review and approval processes.

L'Oreal worked with NVIDIA's Omniverse platform to create 3D models of all its products. Combined with text-to-video prompts, they save countless hours and resources on content production.

Using Generative A.I., we are saving nearly 300% time on generating these product images. - Asmita Dubey , L’Oréal’s Chief Digital & Marketing Officer

Action items for your company:

?? Train and deploy approved A.I. models for each of your brands.

?? Appoint a Gen A.I. lead or owner to work across departments.

?? Establish your A.I. "Code of Ethics" and communicate publicly.


With the rise of Generation A.I. services that interact directly with consumers, artificial intelligence has evolved from being just a component to becoming the main product. L'Oréal is a prime example of a company that embraced this shift by introducing mobile apps driven by AI.

One of their innovations, L’Oréal Paris Beauty Genius??, serves as an A.I. Beauty Companion offering skincare analysis, tailored product suggestions, and personalized skincare guidance through a mobile application. (For further details, check out Campaign US)

L'Oreal Beauty Genius App

Action items for your company:

?? Uncover new revenue sources with A.I. powered products & services.

?? Align business units (corporate venture, marketing, innovation) to expedite.

Generative A.I. is not simply a creative tool for brands and agencies; it's a technology that is ushering in a new wave of creativity and self-expression for your customers. L'Oréal is, again, leading the charge, with partnerships with Ready Player Me , Roblox and Zepeto to allow consumers to generate hairstyles for their avatars like the ones shown below.

This is only just the beginning for Gen A.I. and self-expression between brands and consumers, and we think this will lead to all kinds of exciting new experiences and product customizations designed by the consumer.

Action items for your company:

?? Explore ways of using Gen A.I. to let your customers express themselves

?? Consider how Gen A.I. can be used to help customers personalize products

Picture consulting your "A.I. Employee" for a fresh product formulation that aligns with your target consumer, meets strict business requirements, and delivers the desired product benefits. From drug discovery to product development, Gen A.I. is aiding companies in analyzing data to innovate new products and forecast product effectiveness.

L’Oréal’s Chief Digital & Marketing Officer, Asmita Dubey
"A.I. in general helps in our e-formulation. It's not just about personalization, it's helping with creating new products as well." - Asmita Dubey ,

Businesses are now looking towards internal "data lakes" filled with extensive unstructured data as the upcoming frontier. Generation A.I. is pivotal in deciphering scattered data silos, unifying all organizational data to enhance decision-making processes.

Clara Shih, President at Salesforce AI
80% of data in an organization is unstructured. A.I. solves the problem of mining this data for insights. - Clara Shih - President, Salesforce AI

Action items for your company:

?? Establish data guidelines for who can have access to which data sets.

?? Determine which models are best for achieving your desired outcomes.

?? Run a pilot with CTO, CIO, and R&D support and test performance.


The smartphone might one day be substituted by smart eyeglasses and other wearable technology. Henry Fuchs, a Computer Science Professor at the University of North Carolina, envisions a future where our prescription eyewear will incorporate AR displays, cameras, and various sensors.

On-device compute continues to become more powerful, with both sensor data and A.I. computations becoming possible on-device with the latest processors.

Companies like dotLumen use a combination of A.I. and haptic feedback in its headset to help the visually-impaired navigate, using the same spatial mapping technologies that have been deployed in self-driving cars.

Action items for your company:

?? Experiment with phone-based AR experiences (with our partner Zappar).

?? Consider how A.I. and visual recognition systems can enhance AR utility.

?? Learn more about what's possible with on-device A.I. from Qualcomm.

Gone are the days when robot workers were confined to warehouses and factories. NVIDIA has presented a glimpse into the future where robots can be trained to perform diverse tasks by observing and learning human behaviors through a generalized learning model. This remarkable progress brings us closer to the possibility of humanoid robots being trained and utilized in roles such as fast-food workers, bartenders, chefs, and customer service agents.

NVIDIA's new robotics platform "GR00T" can learn from human behavior, and can be trained to perform a variety of tasks. Learn more in this article from NVIDIA.

Action items for your company:

?? Consider how robots can be used to augment manufacturing and logistics.

?? Workshop with your team to discuss customer-facing roles for robotics.

Digital Twins are transforming warehouses and factories through NVIDIA's Omniverse platform. Picture having conversations directly with your warehouse facilities to inquire about your factory's present condition. An AI agent, equipped with real-time access to sensor and camera data from all your sites, would provide immediate responses.

The future of factories uses Generative A.I. and digital twins not only to monitor facilities, but provide autonomous operation. - Andrew Hsiao , Associate Vice President at Pegatron

Action items for your company:

?? Consider the role that A.I. can play in monitoring your stores & warehouses.

?? Take stock of existing sensor data sets and determine which are best for A.I.

?? Apply an A.I. model "agent" to make this real-time data set queryable.

?? Read more about NVIDIA Omniverse in this case study with BMW


Companies To Know:

NVIDIA featured startups using A.I. to advance productivity from content creation to coding. Here's a roundup of the ones you may want to keep on your radar.

  • Anthropic - A.I. safety and research company that's working to build reliable, interpretable, and steerable AI systems.
  • Bria.AI - A comprehensive solution for Visual Generative A.I. built in the right way for commercial use.
  • Character.AI - Creating revolutionary open-ended conversational applications through breakthrough research.
  • Cohere - Retrieval Augmented Generation (RAG) toolkit allows LLMs to accurately answer questions and solve tasks using enterprise data as the source of truth
  • ElevenLabs - Making content universally accessible in any language & voice.
  • Essential AI - Offering a full-stack of A.I. products that learn to increase productivity by automating time-consuming and monotonous workflows.
  • Getty Images - Offering commercially safe Gen A.I. created imagery.
  • Glean - A.I. powered work assistant across all your company's data.
  • LangChain - Turn your LLMs into reasoning engines that take action.
  • Pegatron - Product development and design manufacturing firm.
  • Synthesia - Create professional instructional A.I. videos from text in 120+ languages. Trusted by Reuters, BBC, Amazon and more.
  • Sakana AI - An “A.I. Lab” for advancements in model creation.?
  • Tabinine AI?- An A.I. coding assistant.
  • Writer - The full-stack generative A.I. platform for enterprises.


Join our Corporate Innovation Council:

Digital experiences are key for your brand to establish a competitive edge, and these innovations are flourishing within the startup ecosystems that King Street Ventures actively participates in worldwide.

Become part of our global community, consisting of senior marketing and innovation executives from 35 Fortune 100 companies. Engage with Founders who are influencing the future of brand experience and marketing. Click here to see our current members.


The information presented in this post serves purely educational purposes and does not signify an approval or recommendation. It is incumbent upon any prospective customer considering the vendors referenced in this article to conduct their own thorough due diligence and evaluation of the vendor, product, or services to align with their specific requirements. It is important to note that KSV maintains no affiliations with the companies alluded to in this article.

Innovating in tech, like Aristotle said, is about finding the wisdom in the craft ??. Be bold, as Musk advises, in creating your path. Let's inspire change together!

回复
Julie Mahoney

Head of Brand & Content Strategy, GSMA; 30 years in digital marketing; Former Founder; Ex-Ogilvy

11 个月

Great takeaways!

Barton Rollert

Growth Strategist | Digital Product Innovation | Applied AI | Tech Partnerships & Development

11 个月

Great Summary - thanks Nick!

要查看或添加评论,请登录

Nick Fuller的更多文章

社区洞察

其他会员也浏览了