In the world of technology and innovation, there is a constant battle between startups and incumbents. Each wave of new technology brings with it the potential for disruption and the opportunity for new players to emerge. However, when it comes to AI, the value has largely gone to the incumbents, leaving startups struggling to make a significant impact.

In the world of technology and innovation, there is a constant battle between startups and incumbents. Each wave of new technology brings with it the potential for disruption and the opportunity for new players to emerge. However, when it comes to AI, the value has largely gone to the incumbents, leaving startups struggling to make a significant impact.

Looking back at previous waves of technological advancements, such as the internet and mobile, we can see a clear pattern. In the first internet wave, the majority of value went to startups like Google, Amazon, and Facebook, with only a small portion captured by incumbents like Microsoft and Apple. The same trend continued with mobile, where incumbents like Apple and Google dominated, but startups like WhatsApp and Uber also managed to capture a significant share of the value.

However, when it comes to AI, the story is different. The big AI applications, such as Google's search algorithms, Facebook's newsfeed and ads, and Amazon's Alexa, have all been developed by incumbents. This raises the question of why startups have struggled to compete in the AI space and what it takes for them to overcome the advantages of incumbents.

To beat an incumbent as a startup, you typically need to either build something significantly better or focus on a new customer segment or distribution moat that the incumbent cannot serve. In other words, you need a 10X better product. Perhaps incumbents have won in the AI space due to their data advantage, but as companies now have access to the broader internet as an initial training set and are using models that work more robustly with smaller data sets, this advantage may be diminishing.

Another reason why incumbents have been successful in the AI space is the nature of the markets startups have targeted. Many prior-wave AI companies directly challenged incumbents or operated in hard markets like education and healthcare, where innovation is often stifled by market structure, regulation, or a lack of understanding of end-user needs. Startups need to identify actual end-user needs and unserved product/markets that will benefit from the latest AI technology.

The current wave of AI technology seems different from previous ones for several reasons. The speed of innovation across various areas is remarkable, making it easier to create products that are 10X better than incumbents'. While GPT-3, a popular AI model, has not yet been widely adopted by startups, a 5-10X better model could create a new ecosystem for startups while also enhancing existing incumbent products. Additionally, there are now infrastructure-centric companies with broad adoption and growing usage, providing startups with access to the necessary technologies and creating more opportunities.

Furthermore, there are specific use cases where AI can have a significant impact. Highly repetitive and highly paid tasks like coding, marketing copy, and website images can benefit from AI-powered workflow tools. The ability to summarize or generate text and images in a high-fidelity way is now possible, opening up new possibilities for product applications.

As organizations scale, the pace of innovation often slows down. The focus shifts from innovation to meeting customer expectations of reliability, efficiency, and reasonable prices. This shift in customer expectations can be a challenge for startups, as they need to balance the need for innovation with the need for predictability and stability.

To overcome the scale vs. speed dilemma, organizations can consider refactoring their code from scratch and spinning off a cash cow to start something new. This approach may involve some initial failures, but with persistence and experience, it can lead to success. Many successful companies, such as Apple, Google, Slack, and Instagram, started with a small team and grew from there.

In conclusion, while incumbents have historically captured the majority of value in the AI space, there are reasons to believe that startups will have a bigger share of AI-generated value this time around. The advancements in AI technology, the presence of infrastructure-centric companies, and the identification of untapped markets and end-user needs all contribute to a more promising future for startups in the AI space.

Actionable advice:

1. Focus on actual end-user needs: To succeed as a startup in the AI space, it is crucial to identify and address the specific needs of end-users. Building products that solve real problems and offer significant value will give startups a competitive edge.

2. Embrace the latest AI technology: Stay updated with the latest advancements in AI technology and leverage them to create products that are 10X better than incumbents'. Look for opportunities where AI can significantly enhance existing workflows or create new possibilities.

3. Balance innovation with predictability: As organizations scale, it becomes important to balance the need for innovation with the need for predictability and reliability. Shipping improvements on a regular schedule and focusing on meeting customer expectations of trust, efficiency, and reasonable prices can help reach a wider audience and make a bigger impact.

The future of AI holds exciting prospects for startups, and with the right approach, they can finally start to realize significant value from their AI-related products. It is an exciting time to be in the AI space, and the potential for innovation and disruption is greater than ever before.