This week, I discuss and explain Code Generations, a concept around tools like transformers that claim to write entire code projects from scratch, but on a surface level. These tools can take a simple text statement like "build an app that reminds me to eat every 2 hours" and transform it into a fully functional program (with bugs and some bad code for now).

⁠Whatspodcasting

Is this the end of software engineers as we know them? Don't worry, I'm not taking sides (yet). This week, instead of delving deep, we're floating on what I call surface-level knowledge about AI code generation.

Here's what we'll cover:

• We'll explore the history of AI code generation, from the early days of chatbots to the rise of machine learning and natural language processing.
• How AI code generators are trained, including the role of open-source projects, data cleaning, and a special type of neural network called an RNN (don't worry, it's not scary!).
• The potential advantages of using generative AI for coding.

Don't forget to check out the description for links: Whatspodcasting

Reference

• https://en.wikipedia.org/wiki/Recurrent_neural_network
• https://www.dhiwise.com/post/how-are-ai-code-generators-revolutionizing-software-development
• https://toloka.ai/blog/history-of-generative-ai/
• https://www.proquest.com/openview/021521f853535306c1c7fff4536e87e3/1?pq-origsite=gscholar&cbl=18750
• https://www.ibm.com/blog/ai-code-generation/
• https://g.co/gemini/share/0d3f408e9a25

I'm Drounz see you next week.