AI is having its big bang moment. In the past weeks, OpenAI released GPT-4; Mozilla launched Mozilla.ai; Bing integrated with Dalle-2 with Bing Images; Adobe released Firefly; GitHub launched Copilot X; Google released Bard; and Facebook launched LLAMA.
I probably missed some because I closed my eyes for 13 seconds.
The hype is real, but the benefits are also real. While AI has been a buzzword thrown around for years, we're finally hitting an inflection point where it actually is useful.
This round of artificial intelligence is powered by large language models, that is, neural networks with billions of weights and inputs. These models are huge and take massive amounts of data to train. The end result is perfectly average; a neural network that has consumed so much information can output average responses over a wide range of topics.
Of course, average is better than 50%, and that typically makes it pretty good.
Remember, AI doesn't solve a problem itself. It's an aid. It makes previously complicated tasks easier. Coding will become easier, designing will become easier, and building products will become easier.
AI doesn't magically eliminate problems. The power needs to be integrated into existing flows to unlock the potential. Which means now is a great time to be solving those problems.