Horizontal AI vs Vertical AI

The tiny startup Windborne Systems announced recently they have outperformed Google’s Gemini in weather prediction.

This being the first year with AI Valentines it has come to light that some are becoming emotionally attached to their AI partner.

Airbnb CEO is hoping its acquisition of GamePlanner.ai back in November will help the company create a more horizontal business model. This model would be more like Amazon or Apple.

Findings from Microsoft have found that hackers are using AI now to improve upon their cyberattacks.

Cohere AI has built and released an AI model for low-resource languages around the globe. This initiative involves over 3,000 independent researchers in 119 countries around the world.

While connections with LLM to external systems increase so do the attack vectors for hackers increases as well.

Google’s Deep Mind can do math problems like an olympiad without human demonstrations.

Munch - takes long-form videos and clips them into short-form videos for IG reels, TikTok or YouTube shorts.

Opus Clip - also will take your long-form video and chop that into bit-size pieces for distribution by IG. YouTube or TikTok as short-form videos

The Algorithm focuses mostly on how the use of AI has shown how AI tools have unfairly harmed minority and women job searchers. The author advocates we move beyond AI job search tools to bring a more human element back into the job search process.

Horizontal vs Verticle AI

Artificial Intelligence (AI) has become an integral part of many industries and applications, but there are two distinct types of AI systems: horizontal and vertical. Horizontal AI is broad-based and versatile, able to perform a wide range of tasks across different domains. Vertical AI, on the other hand, is highly specialized for specific industries or applications, offering optimized accuracy and efficiency. The choice between these two approaches depends on the specific needs of the application and the resources available. Understanding the strengths and weaknesses of both horizontal and vertical AI can help developers and users make informed decisions about which approach to use in different scenarios.

What is Horizontal AI?

Horizontal AI refers to a type of artificial intelligence that can perform a wide range of tasks across different industries and applications. Horizontal AI systems are designed to be versatile and adaptable, capable of performing a variety of tasks regardless of the domain or subject matter.

Some examples:

• Natural language processing (NLP) systems, can perform tasks such as speech recognition, sentiment analysis, and language translation for various industries, from healthcare to finance.

• Computer vision systems can be used for object detection and recognition, facial recognition, and other visual tasks across industries like manufacturing, security, and entertainment.

• General-purpose AI platforms like IBM Watson or Google's TensorFlow can be used for a wide range of applications, from predicting weather patterns to analyzing consumer behavior.

What is Verticle AI?

Vertical AI, also known as specialized AI or domain-specific AI, is a type of artificial intelligence that is designed for a specific application or industry. Unlike horizontal AI, which is more generic and versatile, vertical AI is focused on solving a particular problem or performing a specific task within a certain domain.

Vertical AI systems are usually optimized for the domain they are designed for, which makes them highly accurate and efficient for a particular task. They often use specialized algorithms and training data specific to their domain, which allows them to learn and perform tasks faster and with greater accuracy than more generalized AI systems.

Some examples:

• Medical AI systems, can analyze medical images, diagnose diseases, or predict patient outcomes.

• Financial AI systems, can detect fraud, optimize stock portfolios, or manage risk.

• Manufacturing AI systems, can monitor production lines, predict maintenance needs, or optimize manufacturing processes.

• Transportation AI systems, can control self-driving vehicles, optimize traffic flow, or plan logistics for delivery routes.

I favor the verticle AI model rather than the horizontal AI model. Specialized systems tend to do better than their broader-based counterparts in many different circumstances. You see this especially in AI where small language models or SLMs can produce the same output as an LLM. These SLMs are also more efficient and cost-effective compared to LLMs in practice.