We will need a national AI infrastructure by 2024; read the analysis

In the series on what to expect from artificial intelligence in 2024, Rodrigo Nogueira, of Maritaca AI, explains how Brazil can reduce technological dependence next year

In recent years there have been notable advances in artificial intelligence (AI)especially in large language models (LLM). For example, in 2020, the GPT-3 I answered about 38% of the ENEM questions correctly, not far from random, 20%, in this five-alternative multiple-choice test. In early 2022, GPT-3.5 increased this to 70%, and in 2023 we reached 90% with GPT-4. Most of these advances are due to the use of more interconnected computers and training data, without necessarily relying on scientific breakthroughs or revolutionary ideas. This ability to improve products with investments in hardware rather than human resources for research and development is a unique feature of the field of modern artificial intelligence; No other industrial area benefits so quickly and directly from increased capital-only investments.

This trend has not yet shown signs of exhaustion and it is expected that in 2024 we will have even better AI, which will combine different input and output modalities (images, text, video, sound, etc.) and their use will be increasingly widespread in companies , in public power, in science and among people in general.

There is already evidence that the flaws we have noticed using today’s most powerful AIs, such as the inability to develop a complex, error-free computer program, should soon be fixed. For example, model performance Twins, recently launched by Google, compared human experts on a series of questions from more than 57 knowledge areas, including medicine, mathematics and law. Therefore, it is not difficult to imagine that we are approaching a turning point, when these models will begin to suggest elaborate ideas for decision making, with better results than the proposals put forward by the best human minds. With this possibility imminent, 2024 will be the year in which discussions about artificial intelligence centering on national security issues and its social impact will intensify.

For Brazil, I hope – more than I really believe – that efforts to develop a national AI infrastructure will be intensified. The reason for this effort is similar to the creation of the Companhia Siderúrgica Nacional (CSN) in the 1940s, which accelerated the country’s industrialization and reduced dependence on imported manufactured goods. To be less dependent on technology and more sovereign, we need a similar strategy, namely to be able to develop and deliver our own core AI, including a large-scale IT park, AI training expertise and legislation that supports and encourages domestic industry.

*Rodrigo Nogueira is the founder of the Maritaca AI laboratory

Source: Terra

Rose

Rose James is a Gossipify movie and series reviewer known for her in-depth analysis and unique perspective on the latest releases. With a background in film studies, she provides engaging and informative reviews, and keeps readers up to date with industry trends and emerging talents.