AI and all the hype

T.L; D.R.-> AI is all about the hype, but also a step forward.

I know. It is all hype about A.I. The concept has been around for years. On social media and in the advertising industry for a while now. The number of apps powered by machine learning is increasing by the dozens and some of them now are called AI.

A.I. or Artificial intelligence has been a buzzword for a long time. Take Turing´s essay for example. It does not mean that it is not important or that it will effectively be a new tool at work or not. It will certainly not take our jobs but can make our everyday tasks a lot easier. The hype has endured several news cycles and looks like it will be a buzzword for a little more time. While I wonder if this is the first step in a very long road towards AGI or just another tech marketing stunt, several VIPs signed a letter asking to stop training/using LLM AI until we know we can put some checks or controls around it. Also Italian privacy regulators ban ChatGPT on a GDPR violation claim. Why should we stop? LLM AI is only text, symbolic meaning and tons of math running in a lot of very large rooms filled with servers according to this Microsoft Definition. Not very different from Github Copilot or even Google Search. I do not see the harm.

I have enjoyed Sci-Fi all my life and my soul brightens up with just the idea of AI, autonomous electric cars and advanced robotics. I have written a couple of Sci-fi Short Stories myself imagining how these A.I. will evolve. But all we have seen now with AI is that they are just more advanced search in-text tools where most of the answers have to be verified since most of them are wrong at first glance. For example, Vicky, a bright engineer who is an expert in the ML field and has written a ton around MLs has played with ChatGPT and provided some examples in her blog of how the AI tools are getting it wrong.

While these modern tools called AI are useful, I think they are not still fulfilling the formal A.I. definition and thus, it is a marketing stunt to sell these text tools. But then again is a step forward and a big one since it is getting more monetization options and a lot of funding to get a real A.I. working.

The A.I. definition is very debatable since we still do not have a unified view of what is intelligence and how we can measure it. Then also thinking about self-consciousness, abstract thinking and math-solving skills by a computer is still, in my point of view, a matter of Science Fiction or Philosophy. If we code directly into a machine the ability to be a self-conscious being, identify objects with a camera, or solve math problems we may never know if the system of the machine “understands” the concept of self-being, or “grasps” the multiple meanings of a word that will represent an object through its camera.

What excites me is that small but constant progress toward our technical wonders can extrapolate that we may, someday in the distant future, achieve AI and then AGI. And I believe that there is nothing to fear since AGI by definition does not have anything to do with ethics or morale or even compassion. It will just pursue the most intelligent path there is.