AI at the Forefront: Cutting-Edge Tech and Media Innovations
Chris Pfaff · May 7, 2024
venture capital
(00:00) I was involved with New York Ventures before starting my consultancy. We work primarily in media tech, particularly machine learning and AI for media, OTT, and extended reality. We're always scouting for the next big thing, focusing mostly on Europe and North America.
But way back, I was on the agency side, working with companies like Sony, Sharp, and Yamaha. CES continues to cater to these major tech firms, especially those from Asia. I’ve been attending CES for over 31 years, and this year felt different; it was less about flashy concepts and more about real, scalable innovations.
(00:43) This show had less “smoke and mirrors” and more tangible, incremental innovations at scale, which was exciting for all of us. To give an overview, I’ll run through a few highlights I noted, and then I’ll provide a link for anyone wanting to download the information for further review.
(01:29) One significant brand and AI announcement came from Volkswagen. CES has essentially become a car show, especially for concept cars. Volkswagen’s announcement feels like something that would have seemed incredible six months ago, but now it’s almost expected. It’s interesting how quickly this kind of conversational media technology has become the norm. I still wonder when Amazon, with all its AI expertise, will integrate more advanced AI into its Alexa devices.
(02:44) Speaking of AI, I added something unique to this deck—an image created by DALL-E 4. With a little prompting and fine-tuning, I managed to get it to produce a pretty impressive graphic. Interestingly, some of the major announcements last week didn’t come from CES itself, like the OpenAI GPT store launch. I’m not fully sold on it yet but would love to hear others' opinions. It feels like a lot of people will stick to ChatGPT’s main interface rather than navigating through specialized apps.
(03:55) AI-powered robots were everywhere at CES, but I’m skeptical. Many of these devices seem to add more complexity to home management without actually saving time. For instance, devices like the Roomba are practical because they have a clear purpose—cleaning. But a lot of these new AI devices, like home monitors on wheels, don’t seem to offer much practical benefit.
(04:29) LG showcased “affectionate intelligence”—using AI for consumer monitoring via facial recognition, which could potentially feel intrusive. HiSense also had an interactive art TV, adding a generative AI layer to let people create paintings to display. AI is even in appliances now, like washing machines that detect fabrics and adjust cycles accordingly.
(05:14) Samsung, for example, is releasing an AI-powered phone today. Speaking of new tech, I ordered the Humane AI Pin, just out of curiosity. I might end up returning it, but it will be interesting to demo. Surprisingly, it doesn’t come with a subscription fee, making it more appealing.
(06:37) Amazon, despite Alexa’s current limitations, is working on integrating advanced AI into its devices, partly through partnerships with companies like Character.ai. Walmart also made a big impact this year, promoting its AI-driven “adaptive retail” concept.
(07:17) There were some big announcements in AI customer service as well, like Pfizer’s new AI-enabled help desk, which garnered more attention than many of the CES product launches. Even Apple took the spotlight with its announcements, seemingly overshadowing CES itself.
(08:04) Then there were some oddities, like AI-powered roller skates and a “gen backpack,” both showcased at Eureka Park. Overall, there was so much AI that it felt like an overload. I’ll share the deck link for anyone interested, and I’d love to hear more thoughts from Chris or Michael.
(08:50) Chris, anything stand out to you? For me, I found it interesting to see what big players in e-commerce, like Alibaba and TikTok, were doing. Walmart’s adaptive retailing, while innovative, brought some humorous thoughts to mind—like how we might have called it “shoplifting prevention” back in the day! The Rabbit R1 also got a lot of buzz; it feels like the type of gadget we might have seen only in Asian markets a decade ago.
(09:25) The sheer amount of AI integration in TVs, like TCL’s battle for brightness and enhanced connectivity, was impressive. The Amazon Recognition technology in Alexa, along with AI-enabled apps, has been around for a while, but there were definitely cool integrations at CES.
(10:39) What surprised me was the subtle integration of AI into consumer products. While I expected more brands to showcase partnerships with OpenAI, like the ones we saw at South by Southwest with Shutterstock, it seems much of that was kept behind closed doors.
(11:52) CES had some quieter but effective AI integrations, and I was expecting more bold displays, particularly from XR companies and brands like Meta, which was notably absent in showcasing AI advancements. Even the XR and haptic sections were relatively low-key.
(13:00) Chris, you mentioned the growing automotive presence at CES, now almost resembling an international auto show. Some companies, like Kia and Mitsubishi, presented impressive AI-assisted driving features, showing real advancements in vehicle design and ecosystem connectivity.
(14:02) Kia, in particular, had a great showing with innovative design in AI-powered driving. Their display was less gimmicky and focused on creating a connected ecosystem for the car, which was refreshing. It felt like companies were making real strides toward creating cohesive products rather than one-off tech features.
(14:32) Next year’s CES will be especially interesting. Companies invest heavily in their booths, and we’ll see what advances they bring. There was buzz about “see-through” TVs, and while I’m still not convinced of their practicality, Samsung’s overlay with sports broadcasting graphics was a bit more innovative.
(15:48) The see-through TV by Samsung had an interesting layering effect that allowed for a parallax view of game graphics over actual gameplay footage, creating an AR-like experience. I’m still not sure about the point of see-through TVs, but the Samsung Frame remains a top choice for blending a TV into a home environment.
(16:55) We also saw minimal auto-stereoscopic displays, which allow for 3D viewing without glasses. It feels like we’re close to a breakthrough in 3D display tech that could revolutionize the industry.
(18:12) Sony’s demo of VR production, especially their racing simulators with PSVR2, felt like a good use case for VR. It was surprisingly comfortable and immersive, showcasing VR as an ideal addition to high-end gaming setups.
(19:25) Kia’s modular van, set to launch in the U.S. by 2027, was one of the most fascinating displays. The design feels forward-thinking, like something Apple would create, and hints at a shift towards self-driving, multifunctional vehicles. It’s amazing to think how driving will evolve by the time today’s kids are on the road.
(20:36) Seeing brands emphasize integration of gaming engines like Unreal Engine for car interfaces was a surprising trend. The smaller chip sizes have freed up space in car interiors, making room for enhanced sound systems, displays, and more, which paves the way for exciting design innovations.
(22:03) Amazon also had a small but notable presence, showcasing their work in fleet management and their new automotive embedded Linux distro, highlighting the convergence of automotive and computing technology.
(23:12) AI-driven roller skates were an unusual addition, using AI to learn a person’s gait, which has implications for applications beyond just mobility devices. The potential to integrate this kind of real-time responsiveness into cars or robotics is huge.
(24:27) AI in physical environments is advancing significantly. In building management, for instance, AI is being used to optimize systems in real-time based on user patterns. A fenestration client I work with in the UK has seen how older digital twin technologies are making way for more intelligent, dynamic systems.
(26:06) Sony’s innovations at CES stood out, like their LED panels and integrated virtual production systems. The cohesive ecosystem they’re building between cameras, LED displays, and other hardware is setting a high bar, though their technology remains costly.
(27:23) Sony’s integration of their hardware and software solutions is remarkable. They’re creating an ecosystem where cameras, LED panels, and production tools work seamlessly together, which will likely reduce production costs over time and make high-end equipment more accessible. This democratization will especially benefit content creators who don’t need the ultra-high resolution required for national TV but want high-quality production for platforms like Instagram or TikTok.
(28:03) Nvidia’s customized build of Unreal Engine is another development to watch. They’re embedding AI models to enhance facial animation and clean up motion capture, which could bring significant advancements in visual realism for marketing and entertainment. While much of this tech is still on the experimental side and often requires prompt engineering expertise, the next couple of years should see functional tools emerge that make these capabilities more accessible and intuitive.
(29:22) This diverse expertise is what makes these discussions so engaging. People from various backgrounds bring unique insights into what’s happening in AI, tech, and media. For anyone following CES remotely, feel free to share what stood out to you.
(30:03) I'm curious about thoughts on the Vision Pro. Apple’s entry into the mixed reality space is intriguing, especially given the lack of any major reveal from Google and Microsoft’s pivot to more profitable sectors. As a standalone device, Vision Pro seems ideal for business applications, but its practicality as a consumer product is still up for debate. Apple’s simplicity could give it an edge, especially in B2B, where streamlined setups are critical, but its $3,500 price tag may limit mainstream adoption.
(31:16) For developers, Vision Pro could be revolutionary. It allows for screen mirroring from Mac devices, potentially eliminating one of the biggest headaches in VR development—the constant need to remove and replace a headset to check progress. By streamlining this workflow, Apple could make the device appealing to creators, despite the high price.
(32:37) In AR, simpler glasses like Meta’s new AR-enabled sunglasses are likely to gain traction. I see a strong use case for AR in areas like advertising, where brands could create virtual products for consumers to view in real-world environments. Even Snapchat’s Spectacles hinted at this potential, and Meta’s recent tech demo, including Lex Fridman’s interview using their virtual avatars, shows how realistic these virtual experiences are becoming.
(33:59) There’s a clear trajectory toward using AI-generated assets for brand experiences across platforms. From AI-generated images to full-scale virtual brand environments, companies can now adapt content for platforms like Fortnite’s Unreal Editor, creating interactive, immersive brand installations. This versatility is opening up new possibilities in marketing.
(35:28) For professional settings, Meta Quest is already being used for virtual meetings, though it remains a niche application. The transition from niche tools to broader adoption will likely depend on ease of use and the software ecosystem that supports it.
(36:06) For developers, the Vision Pro’s ability to run Unreal Engine could be significant. Unreal Engine’s continuous improvements and integration with Apple’s hardware will likely make it a powerful tool for both AR and VR applications, though Apple’s ongoing legal issues with Unreal may add complexity to this integration.
(37:29) Vision Pro’s developer potential is intriguing, especially for 3D content creators. While its consumer appeal remains uncertain, its ability to streamline VR workflows could make it valuable as a developer tool. As a consumer product, though, Vision Pro may struggle against Sony’s VR2, which offers high-end VR at a much lower price and is already compatible with popular platforms like PS5.
(39:50) We also saw a strong presence from South Korea this year, particularly with SK’s massive display on sustainability in media emissions. Their eco-conscious initiatives, like reducing file sizes, mark a forward-thinking approach to environmental impact in tech. France, too, had a considerable presence, with investments in emerging technologies like LLMs (large language models). President Macron has been a notable supporter of the tech industry, and French startups are increasingly investing in AI innovations, including language models and other tools aimed at supporting French-speaking users.
(42:02) The presence of media and marketing professionals at CES was unprecedented, with major figures like Linda Yaccarino of X and Mark Cuban speaking at panels. MediaLink hosted their traditional event, and Netflix made a debut appearance with a remarkable booth. These kinds of takeovers highlighted how media and tech are converging.
(44:00) From a media perspective, there was a noticeable divide in CES's atmosphere. In entertainment-focused discussions, people expressed concerns about AI potentially replacing jobs, while on the show floor, there was an optimistic, almost utopian vision of AI-enhanced technology. This contrast in perspectives made for a head-spinning experience, reflecting both the excitement and ethical concerns around AI’s future.
(44:42) It was surprising to see France’s large stake in AI, with Macron’s continued investment in the tech industry and AI language models. France’s focus on non-English language support reflects their desire to make AI accessible and responsible across diverse cultures.
(46:44) Sam Altman recently spoke at Davos, sharing OpenAI’s intentions to prevent the use of its technology in political campaigns. While it’s a step in the right direction, enforcing it effectively may prove challenging, given the complexity of monitoring AI use cases.
(48:07) Casey Newton of Platformer wrote on this recently, acknowledging OpenAI’s efforts to curb election manipulation, even blocking certain tampering requests. His perspective is that OpenAI’s approach could serve as a model for other platforms, though there are still significant challenges ahead as we approach GPT-5.
(48:43) On a separate note, I wanted to share an alternative to Upstream called Mingle, which offers a similar platform for professional networking and virtual events. We’re hosting an introductory call soon for anyone interested in exploring this new technology.
(50:34) A topic that keeps surfacing is AI companionship as a service. This concept goes beyond AI in customer service or sales to more personal realms, raising ethical questions about its potential impact on human relationships. There’s been an influx of requests for virtual AI “companions” that can adapt to individual user preferences, and the implications for this technology are vast, especially as we see developments in AI-driven avatars and conversational agents.
(51:46) AI avatars are quickly becoming more lifelike, and tools like Haen allow users to generate realistic AI-driven virtual companions. Nvidia’s work with facial animation in Unreal Engine is making these avatars look increasingly authentic. But as companies start embedding these avatars in customer-facing applications, there’s a need to address ethical boundaries.
(53:05) The “Bot Love” podcast by Radiotopia delves into the complexities of human relationships with AI, exploring how these connections can be both beneficial and damaging. In some cases, AI can help people manage isolation, but it can also impair real-world relationship skills. Given that AI companionship is inevitable, the focus should be on educating users and creating transparent conversations to avoid unintended consequences, as we’ve seen with social media’s impact on teens.
(54:45) Many clients are interested in creating sales and customer service avatars that look and feel highly realistic. These advancements open up new business possibilities but also highlight ethical concerns around user manipulation and the potential for deceptive practices in consumer interactions.