Skip to content

DSB #143


After a very long pause, I’ve decided that I want DSB back in my life. But everything evolves, including this bulletin. I prefer a more compact format now, so I’ve chosen to include only three categories: Science, Pop, and Joke, to make DSB more readable. For the same reason, I’ve intentionally not included any articles about the recent events at OpenAI. There is now limited space for marketing stunts or AGI nonsense.

And as always, enjoy your reading.

Science – A 1-hour introduction to Large Language Models by Andrej Karpathy. His videos are understandable and a must-watch for everyone, regardless of expertise level. – Four optimisation solutions: a better algorithm, a better data structure, a lower-level system, or a less precise solution. Obvious? Maybe. But also beneficial and useful.
“Premature optimisation might be the root of all evil, but overdue optimisation is the root of all frustration. – Big data with Polars? Easy peasy. Overcome local machine limitations and reduce operational costs. – Even the robust Large Language Models (LLM) aren’t immune to the tricks of adversarial attacks, turning them into puppets.

Pop – Visa is now offering artificial intelligence (AI) advisory services, including actionable insights and recommendations, to help its customers to boost their businesses. – According to a fintech report by Silicon Valley Bank, venture capital investments peaked in 2021. Currently, funding is slowing down, regulatory constraints are increasing, and everything is becoming more complicated.
–  Get ready for the Gen Z in fintech! They are poised to transform the fintech sector just as the Millennials did before them. While this article may appear shallow at first glance, it contains many interesting links and references.
–  2025’s Q1 forecast in Australia: the launch of a digital bank, focused on the booming yet underserved small business sector.


Be First to Comment

Leave a Reply