We are a mere 6 years from the paper by Google called Attention is All You Need1 which outlined the transformer architecture for natural language processing used in many large language models today. 5 years from OpenAi’s first paper which marked the birth of GPT-1. Improving Language Understanding by Generative Pre-Training2. And 3.5 months out from when ChatGPT was introduced to the world on November 30th, 2022. With GPT-4 having been announced two days ago 3, it’s clear the improvement, and adoption of AI is growing at a rapid pace and people have strong feelings about it.
The internet discourse that has proven the loudest come from two camps. One is the anti-AI camp that believes that AI will cause irreparable damage to artists, and other professionals who stand to lose out greatly if AI is allowed to proliferate unchecked. The other is from the AI enthusiasts, who eagerly talk about how life changing the technology is, and how regulating it before it’s had a chance to grow would kill an entire budding industry.
Regardless of where you land on the spectrum it certainly feels like everything is happening faster than expected. In some sense this isn’t true, as work on artificial intelligence has been ongoing since the 1950s 4 But when I saw a friend from college promoting a ChatGPT based real estate app on Facebook, it made me pause and think. The adoption has been fast. The conservative perspective was that this new wave of AI wasn’t good enough to outright replace certain jobs and skills, so it would be used aid them instead. People would use AI to inspire their work or automate the boring stuff that takes away from the creative process. Overtime we’d gradually see certain fields cede more and more ground to newer and more powerful AI. Turns out for a lot of people even the current iteration of AI is good enough. And not just for aiding in the creative process but producing the whole content, and nowhere has it been more eye opening to me than in Streaming, and on YouTube.