A first-of-its-kind gathering of some of the presumably smartest people in A.I. in San Francisco this week decidedly lacked human intelligence and empathy for those who shelled out as much as $1,799 for a VIP All Access Pass. Sponsored by some of the biggest names in big A.I., including Amazon, IBM, and Microsoft, the inaugural GenAI Summit was poorly planned and produced and left some attendees disappointed.
Upon arrival in the morning, at the entrance to San Francisco’s Palace of Fine Arts, a lovely landmark just off the Golden Gate Bridge’s off-ramp, there was a swirly mess of people out front trying to understand how to enter the building. Word spread through the crowd about needing to be in the right line of eight doors tied to the initials of your last name. The 45-minute+ wait to enter was due to the disarray caused by frantic reception workers thumbing through piles of pre-printed badges vs. a more automated option.
Once inside, there were apparently no human workers who could direct eventgoers to where the morning keynote stage was located after asking dozens of them. After deciphering the one posted map of the event, the main pathway back to that theater area was quite the obstacle course, including a robotic dog/humanoid, the Unitree Go2 (“a new creature of embodied A.I.”) that was striking various action poses such as jumping, stretching, shaking hands, and pouncing in the midst of an awed crowd.
After making our way through the obstacle course, which included weaving through parallel lines of people waiting to buy $7 coffees that were “brilliantly” placed in the middle of the main walkway and being misdirected to a line for VIP-only showgoers, we finally made it to the main GPT Stage a few minutes shy of the first 9:30 am keynote. Given the long wait time to get into the venue and the many obstacles to the theater, we scored front-row seats for the talk titled: “Demystifying LLMs with Mistral AI.”
That feeling of having our first small win of the morning did not last long. There was a series of technical glitches that followed: slides that would not advance with the clicker, loud, scratchy walkie-talkie screeching from the VIP entrance located just stage right of the GPT Stage, and difficulty hearing the two co-hosts kick off the event even in the front row. Most of the slowly growing crowd flipped through their phones, annoyed.
The annoyance grew. The talk by Mistral AI’s Dr. Devendra Chaplot was more of an advertisement for his platform than a keynote for audience members to be wowed. He talked about Mistral AI’s open models, portability across the top cloud vendors, and the fact it allegedly has the best performance in terms of value and speed delivered. The one thing I jotted down in my notes was that Mistral AI isn’t available in Portuguese, which struck me as odd, given the amount of A.I. expertise and innovation in Brazil.
Finally, an event bright spot with human intelligence and inspiration.
Ambati talked about the inspiration for founding the A.I. startup that has raised more than $251M in funding in nine rounds due to his wife’s breast cancer diagnosis. He was motivated to harness A.I. to do good — for medical breakthroughs, drug discovery, hurricane prediction, and helping to solve climate change. He believes that A.I. will usher in a new age of abundance. Ambati presented the Sanskrit definition of “data” (the plural of datum) that trains and powers A.I., which means “to give” or “to impact with energy.”
H2O.ai’s mission is to democratize A.I. by building user-friendly tools and fostering collaboration to make A.I. accessible to everyone. Its vision for A.I. is to make a difference for the better in the world, and to create a new abundance in time savings, space (both outer and inner), and to deliver more energy.
In the midst of recent revelations of big A.I. behaving badly — including scraping our data and copyrighted journalism to build the best LLM models, disbanding A.I. safety teams to reduce the risks to humanity — from disinformation to the “Singularity,” and the Scarlett Johansson lawsuit to prevent using “Her” voice for ChatGPT 4o — Ambati’s commitment to “A.I. for All” and his passion for tapping A.I. for good was refreshing.
The following fireside chat led by former WSJ Deputy Tech Editor Don Clark was more riveting than most of the rest of the Day 1 talks due to his human expertise in interviewing technology leaders. Clark interviewed Rodrigo Liang, Co-founder and CEO of SambaNova Systems, a Palo Alto, California-based A.I. innovator that has quickly expanded and raised $1.13 billion in funding across multiple rounds.
The core marketing theme for the fireside chat sponsored by SambaNova tied into the A.I. company’s recent Llama 3 speed record of 1,000 tokens per second. Liang emphasized the importance of this breakthrough to help enterprises fine-tune their private data and his platform’s ability to inference at full 16-bit precision to minimize hallucinations and achieve the world-record performance using only 16 chips.
Of course, Clark — an expert in chip technology — asked Liang why there’s not been more progress and impact from newer upstarts compared to A.I. chip leader NVIDIA. Liang quipped that he and others are catching up to NVIDIA’s 30+ years of innovation, and his company’s focus on fine-tuning architecture to optimize data flow vs. using “rigid” hardware so enterprises can run the same A.I. models across data sources across various cloud vendors or on-premise.
All speeds and feeds and A.I. modeling trends aside, I longed for more inspiration.
Billed as an event that would ignite “the A.I. revolution in the heart of San Francisco,” it left this writer with some lingering questions and concerns. Especially the need to focus on adding more human intelligence, empathy, and purpose into an emerging sector that attracted more than $27 billion in investments to the San Francisco Bay Area in 2023, making it the “undisputed leader in A.I. tech and funding dollars” today.
Funny enough, as I thumbed through my daily print edition of The New York Times (yes, reading print newspapers has become a rarity these days) at lunch, I was struck by two news features on page B3. The first at the top right was a story by Cade Metz about OpenAI training a new model to power its future offerings as it strives towards its ultimate goal of achieving “artificial general intelligence, or A.G.I., a machine that can do anything the human brain can do.” The big A.I. leader also said it was creating a new Safety and Security Committee to “explore the risks posed by the new model and future technologies” – following media reports that most of OpenAI’s former safety team members had been disbanded to other divisions within the well-funded company.
That news piece was juxtaposed with the rest of the front-page Business section feature story on page B3 about A.I. being well-positioned to “replace your C.E.O.” The story by David Streitfeld argues the chief executive has an equal opportunity of being replaced by A.I., just like “the writer of press releases and the customer service representative.” It predicts the potential coming of “dark suites” in corporations to match “dark factories” that are fully automated. Oddly enough, a fall 2023 EdX survey found that nearly half of CEOs are even enthusiastic about that prospect.
All in all, my take is there is a lot of white space in the budding A.I. sector to add back in much more human intelligence, empathy, and inspiration. Even the “flirty” voice of ChatGPT 4o (no longer that of the eerily familiar voice of ScarJo) won’t solve that. What we need is more motivating and persuasive voices, such as H2O.ai’s Sri Satish Ambati.
Chris Knight is a Grit Daily Leadership Network contributor with more than three decades of experience as a journalist, photographer, creative director, copywriter, editor, producer, publicist and PR guru from San Francisco and Silicon Valley. He is the co-founder and creative director of Divino Group and a founding member of The 25 Group.
