African AI researchers are determined that this will change. They are forging their own paths, developing tools that respond to the needs of Africans in their own language.
They face many obstacles. AI research is astronomically expensive, and African startups receive a fraction of the funding that their Western or Asian counterparts do. They must innovate and rely heavily on open-source software to get more done with less.
The African AI story is not just about innovation and persistence, but also about a desire to preserve cultures and influence how AI technologies are used in Africa. Abdullahi Tsanni explains the African AI scene’s opportunities and challenges in this article. He attended the Deep Learning Indaba conference, an annual machine-learning event held in Senegal.
Then, some personal news.This will be my final newsletter. From next week, you’ll be under the capable hands of James O’Donnell. It has been a pleasure to write this newsletter over the past two years. I am grateful that you have joined me on my journey, covering everything from Taylor Swift’s deepfakes to snowballs full of bullshit. I’m still here. I’ll be diving into the AI beat at MIT Technology Review We will bring you stories about what is happening in AI, and how technology is changing our lives and society. Stay tuned for more.
This week, we’re having our biggest sale ever, with 50% off an Annual Subscription to MIT Technology Review. Subscribers receive a digital report on generative AI, the future of work and more. Subscribe here.
Continue reading The Algorithm
Deeper Learning
Why AI could eat the lunch of quantum computing
For years, tech companies have invested billions of dollars in quantum computers. The hope is they’ll revolutionize fields such as finance, drug development, and logistics. These expectations are especially high in physics, chemistry and other fields where quantum mechanics is involved. This is where quantum computers have the potential to be a big advantage over conventional machines.
Enter AI:While the field struggles to deal with the reality of quantum hardware, a new challenger is making progress in some of the most promising applications. AI is being applied to fundamental physics and chemistry as well as materials science, which suggests that quantum computing might not be safe after all.
A growing number of researchers now ask whether AI can solve a significant chunk of the most fascinating problems in chemistry, materials science and physics before large-scale quantum computer become a reality. Edd Gent has more to say about this.
Bits and Bytes
The Saudis plan to build a $100 billion AI superpower
This article examines how Saudi Arabia is looking to get in on the AI action. It’s putting money where its mouth it. The country is investing massive sums to create a tech hub it hopes will rival its neighboring United Arab Emirates. (Bloomberg)
AI makes it harder to tell what is real and not.
Two recent examples demonstrate how AI slop can influence our perception of reality. In Dublin, people gathered around the city centre to watch a Halloween parade. The listing was not planned, but it was created by AI. It was then picked up by local media and social media users. Some social media users, on the other hand, dismissed images of the recent devastating floods in Spain, despite their being real, as AI-generated.
AI companies are becoming more comfortable with offering their technology to the Military
The military has been investing in new technologies including AI. Meta and Anthropic, along with Google and OpenAI, are the newest tech companies to court them. (The Washington Post).
OpenAI has shifted its strategy due to the slowdown in AI tools.
The current paradigm of AI development is to make everything bigger in order to make it better. OpenAI’s Orion model, codenamed Orion performs only slightly better than its predecessors. OpenAI will instead focus on improving models after initial training. (The Information)