Main Topics to Discuss in AI

From Pattern Recognition to Cognition: How far have we truly come in imbuing machines with common sense?

My background as an AI scientist is in commonsense reasoning, which is touted as the "holy grail of AI". I actually got started in this field 18 years ago and the timing for starting my company, Verneek AI, had everything to do with how far we could come in imbuing machines with common sense, which in turn gives rise to all these real-world applications of AI that we are all seeing. You can read more on why we founded Verneek in 2021 of all times, inline with my bio. I decided to pursue real-world applications of my academic work (which is the only true test for the efficacy of AI), and given the consumer-facing applications of my company, I can draw the real picture of how far we have come in building AI with common sense, not based on our academic benchmarks but real-world impact.

Redefining AI: Augmented Intelligence

Alternative title for the session: Redefining (Re)Generative AI

One of the main perceived threats of this overexcitement around AI has been job replacement. I believe in the true power of AI in augmenting us, not replacing us. I believe that we should, collectively, pursue helpful applications of AI that are good for all the stakeholders involved, augmenting humans and augmenting businesses. If I had to redefine AI, I would have called it Augmented Intelligence. Our mission at Verneek is very much inline with this: to build the most helpful AIs that can augment human intelligence anywhere, at anytime. With our field of AI having been widely rebranded as Generative AI overnight, I prefer to ground us in Regenerative AI: using AI as a restorative tool for addressing some of our most existential challenges and improving our existing processes. We are pursuing various such applications of AI at Verneek, that are good for anyone involved.

Domain-general Intelligence, One Domain At a Time

Unlike the mainstream, we believed in deploying domain-general intelligence, one domain at a time! This is unlike the path that the likes of OpenAI and Anthropic are pursuing. I actually believe that the best path forward for building the most helpful AI for the benefit of society is to build AIs with Deep Domain Expertise. Having one general search engine with the average wisdom of the crowd is great, but I believe that the most meaningful value for the sake of all stakeholders will be created by deploying small, controlled, and deep domain expert AIs, right where the pain points are. 

(Un)Sustainability of LLMs for the Future of Big Tech

When we were starting Verneek 2.5 years ago, we thought a lot about the sustainability of building LLMs as the sole service being provided by a startup. Unlike a majority of other deep tech AI startups, we decided that being in the sole business of developing LLMs does not make it into a sustainable standalone business, given its huge dependency on what I call cloud funding -- coming primarily from existing big tech companies. Startups have been historically the main way to bring innovation to the US economy. How can we create the future of big tech, fueled by this tremendous platform shift, that is not already owned by the current big tech incumbents? 

Economy at the Age of Gen-AI: Winners and Losers

With all the board-level conversations around Gen AI strategy of corporate US, where do we see the biggest opportunities across sectors? Who are the touted winners and who are the unlikely top contenders? At Verneek, we are building consumer-facing AI for enterprises across various sectors, so I have insights into the evolving board-level conversations at various organizations pre- and post-ChatGPT, and have a particular take on the unlikely sectors that I see as ripe for disruption. 

Content Creation in the Age of AI

With all the Online content getting freely scraped by the LLMs and leaving out expert content creators, Media companies and content platforms have two choices: to either wait and see their value getting wiped out and their content getting monetized without them in the loop, or put their content behind registration wall and spin up their own AI for better engaging with their audience! The latter is what Verneek enables media and content platforms to do with a turn-key solution, specially those that don't have any internal AI teams to build it in-house.