When Google arrived, at the end of the 90s, the amazing thing for us who were users of the first, chaotic Web, were the response times: on a clean white screen there was, and still is, a string where you could write a request and when the response arrived there was ‘it was written that it had reached you in a certain number of milliseconds. The fact that our application could arrive on Google’s servers to California and back with a sensible and useful response in that amount of time that we can compare to the blink of an eye, it was magic.
At the time, the rule on the Web, when you wanted to browse, was to wait: even just waiting It took a good minute to go online in which the computer returned sounds that many became fond of; and loading an entire site often required a lot of patience. Since then things have changed a lot, thanks to ultra-broadband and PC processors, everything has become faster and more fluid. One thing was missing: the possibility of conversing with an artificial intelligence with human reaction times. If a chatbot or robot takes too long to respond, they immediately reveal their nature, the fact that they are made of silicon, of not being human. As if I felt the wheels turning.
The new version of ChatGPT presented yesterday, in addition to being able to interact with sounds, photos and videos, as well as text, makes the wheels disappear, i.e. shortens reaction times even further. Which now ranges from 200 to 500 milliseconds to analyze the request, understand the context and generate an appropriate response and transmit it from the server to our computer. Half a second at most, or, in the simplest cases, half of a half second. Faster, on average, than a human reaction. It may seem like nothing but it’s a lot: we are creating artificial tools capable of perfectly understanding our language, therefore having a precise idea of how the world works, and communicating with a promptness similar to that of people.
If this thing has the impact that Google has had on our lives, these agents will soon be our personal assistants that we will talk to all day long. Will be However, it is important to remember that they are not humanwhich are just mathematical models trained to give great answers.