Design by Artis Briedis
The similarities with the Internet development curve are striking – we have just passed the point where GenAI has been called a bubble and are approaching the stage where everybody is either too busy with other priorities to do anything or has decided to react and build an internal tool of their own. It is going to take some time still to arrive at a point when thousands of companies and their business models will get disrupted allowing many new businesses to be created.
This is the core message from Aditya Challapally, a teaching instructor at Stanford, who has done some recent research on the adoption of GenAI. These are my three takeaways after watching his webinar on YouTube:
Lesson #1
The statement about data being the new oil is not true any longer. The latest versions of large language models are already very good out of box which has moved the main focus towards distribution.
How to get the product into the hands of as many as possible and as quickly as possible now has become the key issue. Distribution which has been transformed into the main moat offers the non-tech players a unique opportunity to compete against the tech companies. Non-techs have different users than techs which provides them with a significant advantage.
The balance of power has moved away from those who had talent, infrastructure and models towards those who have better distribution, superior user experience and data. Research suggests that people will even accept a slightly worse user experience if the results do help them and provide value.
Lesson #2
Hardware is likely to remain a very lucrative, durable and defensible market niche. Contrary to popular belief model creators are not the ones who are making or will make the most money. It is reasonable to expect that only a few companies will remain in this space.
The best opportunities are to be found in mid-tier to large non-tech companies that have troves of data and a large existing user base. The real value can be unlocked by integrating GenAI to enhance the existing systems. Even the most basic AI functionality when implemented seamlessly into an existing website or app can create lots of value.
Lesson #3
The next iteration of foundation models usually beats the customised previous versions (with a few exceptions only).
The best approach for companies is to take the model out of box and find a user scenario that somehow works. It is critically important to always build a user-facing feature.
For startups, the good news is the paradigm shift away from money, models and thousands of employees to user experience (UX) and user interface(UI), unique data and ten employees at most. Going against established companies that are moving slowly or not moving at all offers an opportunity to disrupt whole markets.
And, as a final note, – Aditya’s research suggests that users prefer products that help them (for example, to summarise texts or emails) and are not so keen on using products for creation which can perform all the work for them. And almost all of the respondents do hate chatbots!
For those willing to try and pick top lessons of their own here is a link to the video.
* from anything that you are reading, watching or hearing you can realistically expect to remember only a limited number of things. My solution is to pick just 3 items or ideas from any material. This number is non-negotiable. Even the most extraordinary experience gets compressed into 3 things to remember. This approach has worked well for me.
This note was first published on Medium on 16 November 2024.
Aivars Jurcans has more than 20 years of corporate finance and investment banking experience. His services are currently available through Murinus Advisers. More of Aivars’ writings can be found on his page Corporate Financier’s Notes.
