The Wheat from the Chaff #2

As part of my job, every week I read a lot of content about data, technology, and artificial intelligence (also known as AI). I filter them and send you the best ones, every Sunday.

Hello everyone 👋 ,

This week, we had some important updates! The proposal to regulate the use of artificial intelligence (AI) in Brazil is already in the Senate. Additionally, it's important to remember that data is a crucial element in the revolutions we are currently witnessing. Software has been a major wave of innovation that has radically changed many things, but will there be a next wave?

It is time to separate the wheat from the chaff…

Less is More

Data is one of the main inputs, if not the most important, for all the solutions and advancements in AI that we are witnessing. In fact, data plays a fundamental role, but for a long time, it has been a barrier to the evolution of the technology itself and, subsequently, to its adoption in certain sectors. However, this is changing rapidly.

In more regulated sectors, such as finance, some initiatives are being led by the government itself. An example of this is open finance. Digitalization driven by technology plays an important role in this context. In a nutshell, there are 4 main types of data: tabular data (like those in your spreadsheet), images, text, and voice.

According to cloud storage company Domo, in 2022, users of the YouTube platform were uploading 500 hours of video per minute. In April 2022, the internet reached 63% of the global population, which represents approximately 5 billion people. Among these users, more than 93% - that is, 4.65 billion people - use social media. According to Statisa, the total amount of data to be created, captured, copied, and consumed worldwide in 2022 is 97 zettabytes, and it is expected to grow to 181 zettabytes by 2025.

Understanding the data landscape, a question arises: how do data relate to all this buzz around ChatGPT and similar technologies, as well as to my career on business?

Taking a step back, the ultimate quest is for automation that impact revenue or cost, and AI can certainly support this process. In automation, some data is consumed, processed, and understood in this context, resulting in a suggested action.

Regarding text data, advancements began around 2016, making them quite recent.

​​Companies and products have many processes involving text in tasks that currently have little automation, such as summarizing meetings, writing responses, recording opinions, etc. However, existing AI technologies did not perform well in this context. How many chatbots have you tried to use without resolving anything?

Hence, the arrival of ChatGPT, Bard, and similar technologies is extremely important. They are generic language models that understand natural language and will be used to develop thousands of products and technologies involving text processing for decision-making and task automation. Now, the race is on to automate more specific tasks based on processes that involve text as basic information.

Interesting topics:

• 🔗 Link: How “Reclame Aqui”, using Natural Language Processing (NLP), managed to reduce moderation efforts and response time to complaints.

• 🔗 Link: A Bill is seeking to regulate the use of artificial intelligence in Brazil.

Trending topics:

• 🔗 Link: The software-driven revolution has indeed transformed the way business is conducted. And the next wave of innovation is expected to be automated processes that continuously improve based on artificial intelligence models, without the need for constant human intervention.

Guest of the Week: Dall-E

Every week, we will bring a tip from a guest. This week, we challenged Dall-E, an AI program that creates images based on textual descriptions, to tell us how it sees Carnival in Brazil. Here's the response:

See you Sunday…

We always arrive in your inbox around 12:08 PM. Some email servers are stubborn and delayed... Others are even worse and throw us into spam and/or promotions. What's the logic behind this timing? ...None :)