Peter Erickson wears many community-building hats: founder of Modev (conferences and special events for developers), founder of ExoLeaders.com (for executives), and founder of VOICE & AI — an organization whose conferences and events bring together developers, buyers, sellers, and manufacturers seeking to gain entry into this fast-moving segment of the AI market. (VOICE & AI’s next event will be on October 28-30, 2024 in Arlington, Virginia).
Peter has been in the world of computer-based voice search for a long time and is a knowledgeable observer of the sector. While voice search is a category that’s been with us for some time, it’s only recently that its potential to be truly useful is arriving, thanks to the advent of enabling AI technologies and LLMs.
As we note in our talk, the marketing implications of the rise of AI-powered voice search are only beginning to be appreciated, with single-result search engines such as Perplexity constituting “winner take all” environments where the monetization rules are very different from those of multiple-result search engines.
Peter and I cover a lot of ground in our chat; topics include:
- The degree to which innovation in LLM-driven Voice Search will progress ahead of text-based search, given there is less risk for its monetization model.
- How LLM-powered search engines and voice search applications create a “winner take all” scenario for marketers (and what marketers can do about this).
- Why Perplexity is in a good position to innovate in Voice Search, as they have reimagined search and presented it in an organized and addictive way for users.
- The degree to which open source and proprietary technologies will coexist in the industry.
- The high costs of AI-powered voice search and the degree to which costs may moderate in the near future. (example of a network of gaming computers to perform inference tasks at pennies on the dollar compared to traditional models).
- Speculation as to whether Perplexity will be acquired this year (and by whom).
- The costs and benefits of AI systems in education, given that they may “dumb down” languages and dialects.
- The potential social impact of AI on language and dialects usage, particularly in lesser developed countries.
- The potential for AI-powered chatbots to be used for nefarious purposes, such as scams or hacking.
- Whether it is realistic to expect that AI chatbots can provide a meritocratic experience, given the potential for hacking and manipulation.
- The need for the industry to work to address trust issues with AI chatbots; in lieu of this, government regulation may unfortunately be necessary.
- The challenges of programming technology conferences in advance when the technology moves ahead so quickly.
