Major shift for South African software developers

The number of software developers based in South Africa working on AI-based open source projects has boomed in the past year, GitHub CEO Thomas Dohmke has told MyBroadband.

According to the developer platform’s latest “Octoverse” report, South Africa had one of Africa’s fastest-growing GitHub developer communities.

“Nigeria, Egypt, South Africa, and Kenya are considered Africa’s ‘big four’ countries focusing on technical literacy, drawing global investors, and securing most of Africa’s startup funding in 2023,” GitHub said.

GitHub is a developer platform based on the Git version control system that lets developers create, store, manage and share their code.

It reported that over 664,000 developers from South Africa joined the platform in the past year, representing a 23% year-over-year growth.

Dohmke said the growth of AI projects on GitHub in South Africa was nearly double that percentage, at around 45% year over year.

“In the tech industry, and even in the global news, when we’re thinking about AI we often talk about the big companies that are developing the next generation of large language models,” Dohmke said.

“Obviously, there are not that many of those companies, given that model training is very expensive and getting access to such large GPU clusters requires a lot of capital.”

However, while these major AI model advancements are happening, many interesting applications are being built on top of this new technology.

Dohmke said GitHub is also home to many open source components that play a role in AI.

“These open source components play a role in AI too, whether it’s powering the models and the model stack.”

Dohmke said the growth of AI projects on GitHub from South Africa shows that AI is at the top of many developers’ minds, and that they see a lot of opportunities there.

One example of a local AI-based project is the University of the Witwatersrand’s “AI_r”.

It combines an Internet of Things sensor network with AI-based analysis in the hopes of building an affordable system to measure air quality across South Africa.

Ai_r is made up of a collection of small boxes that cost about R1,900 each to make.

These boxes can be mounted on a window sill of any building, where they take air samples and feed this data back to a cloud in real-time.

Modelling and forecasts are made with artificial intelligence using data that is uploaded to the cloud.

Artificial intelligence is used to integrate sets of data, learn from them, and create models that, according to the researchers, save a tremendous amount of resources.

They said in air quality monitoring, artificial intelligence will learn from the vast wealth of data that these sensors produce to make predictions.

“For example, Ai_r will be able to tell us the impact of weather changes on air quality so that we know which areas will have more polluted air,” said Wits physics professor Bruce Mellado.

“Without artificial intelligence, it would be very difficult to come up with a cost-effective system that could monitor air quality and give us predictions.”

Industry analyst Arthur Goldstuck believes artificial intelligence should be viewed as a tool rather than something invasive and disruptive.

Regarding people’s fear of being made obsolete, Goldstuck pointed to the example of Ford’s assembly plant in Silverton, Pretoria, which spent R15.8 billion to upgrade and automate the factory.

The plant acquired 493 robots to transform stamped body parts into a complete Ford Ranger.

However, this did not cut jobs but boosted production and increased the factory’s headcount by 1,200 people.

Even when workers are made redundant by automation, production is often increased, which allows businesses to hire more employees.