#19 Yariv Adan

An absolute blast of an episode with Yariv Adan as he joined me to talk about generative AI technology, driving insights from his career at the center of AI development in a perfect combination with a clear view of the future of the technology. The episode is entertaining, fast paced, enlightening and highly engaging and an I am very happy to publish it for the audience today.

Read more about Yariv at the bottom of this newsletter.

As always, I am very happy that we once again managed to make an episode that explores more than it explains ??

Listen to the episode here:

Apple: The Only Constant on Apple Podcasts

Spotify: The Only Constant | Podcast on Spotify

Google: The Only Constant ( google.com )

Spreaker: The Only Constant ( spreaker.com )

What made this episode special to me?:

Harder to say what didn't. Yariv has been at the center of AI development for so long - having been a central part of the team at Google that developed Google Assistant and also having lead conversational AI in Google Cloud until recently - and having him on the podcast with his energy and drive was a great pleasure for me. In post-production, I loved every second of this conversation, and I hope everyone else will too.

Yariv also becomes a distinctly different voice on the podcast. As I always probe and try to get people to expose some insecurities about the wide ranging hype evangelist reports, Yariv managed to keep a firm and direct focus on what the technology is actually offering, how it differs from previous tech and how he is an ardent believer in the benefits of it - even saying we should be nicer to it. And his arguments are compelling and reminds me that this it is definitely not the fault of the tech if we misuse it - that is, as with all agency, on us.

My other key post-recording deliberations are:

• We are used to precise systems - we expect correct answers. But the thing that triggered us to call this a novel breakthrough, that it speaks like a human, should be assumed and taken seriously: we need to ask it for examples and sources, just like we would of any human. This way, it really is an intern more than an AI overlord...
• On that note: the breakthrough concept of "Generative AI" is what Google didn't see and why they were beat to it by OpenAI. Google was looking for perfect answers, like in its search engine, but the way the entire world just accepted flawed, hallucinated responses that needed review simply caught them by surprise. And as he says on the pod, it is like when one kid at kindergarten starts misbehaving, then everyone does and now all the big tech companies are making unreliable AI models.
• Yariv notes that the team behind Google Assistant might have overestimated the significance of transcribing and mistaken that for understanding - much like probably happened with Siri. That is clear now that we have models capable of understanding and it makes these older chatbots seem monolithic. Makes me wonder - what capabilities are we ascribing to the current GenAI models that isn't there and which in hindsight will look absolutely silly?

And I have more... this is an episode I will be returning to, and I hope you will as well. Enjoy the listen and make sure to follow Yariv and his future journey which will surely be exciting.

About Yariv Adan :

Yariv Adan began his career as a software engineer in Israel in the late 90s. He worked as both an engineer and an engineering manager at several startups until 2007, when he transitioned to product management at Google. There, he was part of the founding teams for multiple strategic areas, including privacy in 2007 and emerging markets in 2009. By 2015, Adan's interest in AI led him to join the small founding team of what would evolve into the Google Assistant and Google Lens. He spent seven years working on cutting-edge AI consumer applications before moving to Google Cloud, where he led Product for conversational AI and, starting in 2022, applied Generative AI.

After 17 years at Google, Adan recently left the company to start a venture capital firm specializing in early-stage AI startups, collaborating with a group of long-time colleagues. He is also actively involved in projects related to AI art and AI-augmented food.