TikTok and the Sorting Hat — Remains of the Day: Early Days of AI

TikTok and the Sorting Hat — Remains of the Day: Early Days of AI

In the world of technology and entertainment, there are two significant developments that have captured the attention of users and experts alike: TikTok's algorithm and the emergence of AI. While these may seem like unrelated topics, there are actually some intriguing connections between them.

Let's start by exploring TikTok's algorithm, which has been hailed as the most important piece of technology introduced by Bytedance to the platform. The updated For You Page feed algorithm revolutionized the way users discover content on TikTok. Unlike other social networks where you have to follow specific individuals, TikTok's algorithm sorts users into various subcultures based on their interests. This means that users are constantly exposed to content that aligns with their preferences, creating a truly personalized experience.

But how does this relate to AI? Well, TikTok's algorithm is a prime example of how machine learning can overcome cultural barriers. By analyzing user behavior and preferences, the algorithm can effectively connect videos with the audiences that will appreciate them the most. This is a stark contrast to other social networks that rely on social graphs, where you have to follow and friend others to build your network. TikTok's interest-based approach allows for rapid and efficient matchmaking between content and users.

The success of TikTok's algorithm can be attributed to its focus on entertainment rather than social capital. While social networks like Facebook thrive on the connections between individuals, TikTok is all about providing engaging and entertaining content. In fact, many users on TikTok don't even socialize with each other or know each other personally. Instead, they come to TikTok for the sheer entertainment value it offers.

This shift from a social graph to an interest graph has been a game-changer for TikTok. Social graphs often suffer from negative network effects, as users are rarely interested in everything from a single person they follow. In contrast, TikTok's interest graph, built around short video content, allows for efficient personalization without burdening the user. The algorithm learns and adapts through the consumption of content, making it a passive and seamless experience for users.

So, what can we learn from TikTok's success and its algorithm? Here are three actionable insights:

1. Focus on user interests: Instead of relying solely on social connections, consider building an interest-based network that caters to users' specific preferences. By providing personalized content, you can create a more engaging and satisfying user experience.

2. Embrace machine learning: Invest in AI technologies that can analyze user behavior and preferences to deliver tailored recommendations. Machine learning algorithms can help overcome cultural barriers and connect users with the content they'll love.

3. Prioritize entertainment value: If you're building an entertainment-focused platform, prioritize the creation of engaging and entertaining content. Users come for the entertainment, so make sure you deliver an endless stream of captivating experiences.

Now, let's shift our focus to the early days of AI. Many people view AI as a continuum of past capabilities, but it's essential to recognize that we're entering a new era of technology. This new era brings about a step function in capabilities and products enabled by AI, marking a significant discontinuity from the past.

One of the key developments in this new era is the launch of ChatGPT, which mainstreamed the idea that AI is a game-changer. However, we are still far from reaching peak AI usage or peak AI hype. Large enterprise planning cycles and the time required for prototyping and building mean that AI's full potential is yet to be realized.

In these early days of AI, we can identify four waves of development:

Wave 1: GenAI native companies. These are companies like ChatGPT, Midjourney, Character.AI, and others that have gained significant revenue and user traction.

Wave 2: Early startup adopters and fast mid-market incumbents. This wave consists of startups that have launched on top of advanced AI models like GPT-3.5/4. Additionally, a small number of multi-billion dollar companies have quickly adopted AI-powered products.

Wave 3: The next wave of startups that are currently being founded. This wave is expected to explore new formats like voice and video, as well as utilizing natural language in various verticals. Companies like Eleven Labs, LMNT, and LFG Labs will contribute to incremental advancements in AI experiences.

Wave 4: The first big enterprise adopters. Due to longer planning and build cycles, larger companies will likely start releasing fully developed AI products in the coming years. This wave will showcase the true potential of AI in enterprise settings.

It's worth noting that Google's MedPaLM2 model has already demonstrated the incredible capabilities of AI in the medical field. The model outperforms human physicians to such an extent that even having medical experts review and improve it can make it worse. This highlights the immense potential of AI in transforming industries and disrupting traditional practices.

In conclusion, the success of TikTok's algorithm and the early days of AI both offer valuable insights for businesses and entrepreneurs. By embracing interest-based networks, leveraging machine learning algorithms, and prioritizing entertainment value, you can create engaging platforms that resonate with users. Additionally, recognizing the transformative power of AI and its distinct era can help you stay ahead of the curve and identify opportunities for innovation.

So, whether you're building the next TikTok or exploring the possibilities of AI, remember to keep these actionable insights in mind. The future of entertainment and technology is evolving rapidly, and it's up to you to seize the opportunities that lie ahead.