Alex Douzet is CEO of Pumpkin, a pet insurance and wellness care provider founded to help ensure pets live their healthiest lives.
Any business leader following the emergence of generative AI is wondering how it will disrupt the current architecture, user experience and monetization of today’s web. Recently, Rich Barton, CEO of Zillow, described this disruption and re-architecture of the web as “the race to intimacy.”
Barton argues that AI is transforming how we access and consume information from a path shaped and informed by the data of the masses and advertisers to one that is intimate—more individual and private. Likewise, it will shift the user experience of the web from web pages and applications optimized for information retrieval and discovery towards intuitive, conversational interfaces with an AI agent.
In this emerging technology landscape, our individual informational needs will shape the online experiences and the content we connect with; search engines will come to seem clunky and suspicious. Information retrieval that looks like a library index is a byproduct of a soon-to-be legacy architecture of the web. It is different from what users want or where the internet is heading. David Friedberg recently explained on his All-In Podcast, “You no longer need an index. You need ‘someone’ to retrieve the relevant information for you.” That “someone” is AI.
Friedberg describes the reverse architecture of the client-server model. Today we are a client, a node on the network, and we communicate with the center of the network. The future will be the opposite. When more of our data about us is aggregated, we become the server.
Imagine a world where every individual has an IP address. Behind our address is our proprietary, aggregated data that we can rent or provide to the brands and services we find relevant, valuable and trustworthy. We see this transformation already taking place today with content creators on Reddit and influencers on TikTok, YouTube or Instagram. Most recently, soccer star Lionel Messi’s groundbreaking Inter-Miami contract (paywall), which entitles him to a piece of every ticket sales and revenue share deal with Apple MLS TV subscription broadcasting, proves that the individual now has power over the network.
A few people believe this AI race will result in a world where vertical problems are finally addressed horizontally. This use case may reach some scale but won’t be as dominant as the natural monopoly that Google achieved with search. Consumers have been burnt by giving their data to a single platform or agent. More than ever, I believe users will resist giving all their data to a singular product or algorithm that creates silos of extremes we’re forced to live within (e.g., Google, Meta.)
Likewise, I find that people are far less likely to have a trusted “relationship” with a single AI or search engine. In my opinion, the fragmentation of deep expertise is more likely to lead to individuals habitually leveraging a select suite of trusted AI agents while periodically interacting with a few broad, encyclopedic agents like ChatGPT and Google Bard. Rich Barton’s vision of “the race to intimacy” describes the need for brands to be highly trusted. Vertical AI experts can converse and understand an individual user’s needs and values.
According to IDC FutureScape 2023 Report, the size of the digital economy in 2023 is expected to be $17.7 trillion. This is a compound annual growth rate (CAGR) of 7.3% from 2022. Google, Apple, Microsoft, Amazon and Meta dominate the tech industry, but the AI race will likely re-distribute that pie. The top of the funnel is up for grabs. I predict that big tech will be a player at the forefront of the AI race for the next ten years to claim their fair share, but they will likely face an innovator dilemma as they shift their focus and resources away from their core businesses.
The cost of building Large Language Models (LLM) will contribute toward verticalization and more specialized AI applications and agents. Today, it is costly to run and train an LLM for generative AI such as GPT. OpenAI raised $10 billion from Microsoft last year, and a few weeks ago, the European company, Mistral AI, raised a seed round of €105 million at a €240 million value. These rounds are significant because running and training an LLM requires companies to purchase a ton of extremely expensive H100 GPUs (Graphics Processing Units).
However, as evidenced by OpenAI lowering the price of its API service, that cost curve is dropping quickly. David Friedberg argues that the cost of running such a model is dropping at ~100X over 12-24 months. According to Friedberg, OpenAI spent ~$400 million to train GPT 4.0. In 2025, the cost to train the same model might be about $4 million to $5 million.
If the very high entry barriers today will be very low tomorrow, there is likely minimal first-mover advantage or gain in being the first to train the model. When technology becomes cheap enough, I believe we will see all kinds of new apps and products developing. Innovative, Vertical AI agents that offer unique, highly-expert services will be the ones that users entrust with their personal information. The real benefit will be a company’s data set, not a cost advantage. It is the classic challenge of venture investing—it’s all about timing.
The speed at which AI is evolving is head-spinning and will likely disrupt every industry and the web as we know it. The resulting “race towards intimacy” will fundamentally change our relationship with technology. The re-architecture of the web and the dominance of conversational interfaces will put users at the center of the information around them—empowering them to retrieve the exact information they want—not what the masses or advertisers think they want. Individuals will demand excellent information, more human and conversational user experiences and far more control of their data and privacy. This won’t come from ChatGPT, Bard or another big tech mega bot. It will come from hundreds of entrepreneurs in hundreds of different verticals that capitalize on the shrinking barriers of entry to create highly intimate AI tools that users love.
Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
Read the full article here