The brief history of artificial intelligence: The world has changed fast – what might be next?

The brief history of artificial intelligence: The world has changed fast – what might be next?

Artificial Intelligence (AI) has come a long way in just a few decades. Ten years ago, machines were nowhere close to providing language or image recognition at a human level. However, AI systems have steadily become more capable, and they are now surpassing humans in tests across various domains. This progress can be attributed to three fundamental factors: training computation, algorithms, and input data.

Training computation refers to the computational power required to train AI systems that rely on machine learning. For the first six decades, training computation increased in line with Moore's Law, doubling roughly every 20 months. However, since 2010, this exponential growth has accelerated even further, with a doubling time of just about 6 months. This increased computational power has played a significant role in enhancing the capabilities of AI systems.

Alongside training computation, the algorithms used in AI systems have also evolved. Researchers and scientists have developed more sophisticated algorithms that are better equipped to handle complex tasks. These algorithms enable AI systems to process and analyze vast amounts of data, leading to improved performance and accuracy.

The third factor driving the capabilities of AI systems is the input data used for training. AI systems learn from large datasets, and the quality and diversity of these datasets greatly impact their performance. As researchers gather and curate more extensive and more diverse datasets, AI systems become more adept at understanding and interpreting various types of information.

Looking ahead, many AI experts believe that there is a real chance of developing human-level artificial intelligence within the next few decades. Some even speculate that it could happen much sooner than expected. In her latest update, Cotra estimated a 50% probability of developing "transformative AI" by the year 2040, less than two decades from now. This prediction highlights the rapid progress and potential future advancements in the field of AI.

The Network Effects Manual: 13 Different Network Effects (and counting)

Network effects play a crucial role in creating defensibility in the digital world. Companies that incorporate strong network effects into their core business models tend to emerge as winners in the industry. A recent three-year study revealed that network effects are responsible for 70% of the value created by tech companies since the advent of the Internet in 1994.

Unlike viral effects, which focus on getting new users for free, network effects are about creating defensibility. The strongest type of network effects is direct network effects, where increased usage of a product leads to a direct increase in its value to users. This direct correlation between usage and value makes it challenging for competitors to replicate the same level of value for their users.

Reed's Law, a concept coined by David P. Reed, suggests that the value of a network increases exponentially in proportion to the number of users. This exponential growth surpasses what Metcalfe's Law describes. As a result, networks with a larger user base have a significant advantage over smaller networks.

Network effects become even more powerful when they are combined with personal utility networks. These networks provide practical utility to users and are typically used for private communication rather than public communication. When individuals see that people they know from the real world are using the same network, it adds value and incentivizes them to join.

However, network effects are not without their vulnerabilities. Same-side users, or users on the same side of the network, can often subtract value from each other. This is particularly true in marketplace networks, where aggregating competing sellers in one location allows them to gain more business. To break apart such networks, a better value proposition for both buyers and sellers must be provided simultaneously.

Additionally, multi-tenanting poses a challenge for both marketplace and platform networks. Multi-tenanting occurs when users engage with multiple platforms simultaneously. For marketplaces, the goal is to add enough value and lock-in to prevent members from multi-tenanting. Similarly, platforms must create sufficient value to discourage developers from creating versions of their apps for competing platforms.

While network effects are a strong defensibility mechanism, they can be vulnerable to competition. Asymptotic marketplaces, for example, are more susceptible to competition as a competitor might be able to provide a comparable service with fewer resources. Data network effects, which occur when a product's value increases with more data, can also face challenges if the relationship between usage and useful new data gathered is not balanced.

Technological advantages, although initially valuable, have a short half-life in terms of defensibility. However, tech performance network effects provide a runaway advantage for products that are the first to market. These network effects arise from the performance improvements of a technology over time, making it difficult for competitors to catch up.

Social network effects, which operate through psychology and interactions between people, are often the hardest to deploy for long-term defensibility. However, successfully leveraging psychology in favor of a product or company can offer a significant advantage. The winner-take-most tendency observed in language throughout history highlights the power of social network effects. Startups can utilize the network effects of language to their advantage by creating business category language and naming their company or product strategically.

In conclusion, both artificial intelligence and network effects have transformed the digital landscape in significant ways. AI systems have become increasingly capable, surpassing human performance in various domains. Network effects, on the other hand, have proven to be a powerful defensibility mechanism for tech companies, accounting for a substantial portion of value creation. Understanding and harnessing the potential of these two forces can lead to transformative advancements and success in the digital age.

Actionable Advice:

1. Embrace AI: As AI continues to evolve and advance, it is crucial for businesses and individuals to embrace its potential. By incorporating AI technologies into operations and decision-making processes, organizations can gain a competitive edge and unlock new opportunities.
2. Build Strong Network Effects: For companies operating in the digital realm, building strong network effects should be a priority. By focusing on direct network effects, personal utility networks, and leveraging various types of network effects strategically, organizations can create defensibility and achieve sustainable growth.
3. Stay Ahead of the Curve: In both AI and network effects, staying ahead of the curve is essential. Continuously monitoring advancements in AI technologies and understanding emerging network effects can help businesses adapt and thrive in a rapidly evolving digital landscape.