The world economy is preparing to pay the bill for a bubble in the IT sector, the "metaverse" is so aborted that soon Pillon will want it outlawed, cryptocoins are now synonymous with fraud, and we are already on the next hype. There is little to do, journalism that is the servant of finance, which I will call timojournalism, does not understand the lesson.
I'm referring to bullshit like this:
I chose this one because it's the most ridiculous, but you'll have to agree that if we're going to indicate any hype, ChatGPT is certainly one of them.
The second is just a blatant rubbish, appeared on twitter, and happens to be the guy is a web designer and happens to explain how to earn money using things that a web designer can do. If he had told him “go and buy truffles at the Alba auction spending 100 euros per gram, and then resell them in Frankfurt where they cost 300 euros per gram”, I would have liked to see him.
Moreover, the "value" of $25,000 was estimated using a tool that makes spanometric estimates, but we would like to see the $25,000.
Why this catastrophic overestimation?
Well, the first is that if you are a financial group that owns La Stampa, Repubblica and Corriere, you will be able to benefit from a flow of people who call their stock broker and say "buy Microsoft". Just to say, the effects of ChatGPT on the stock exchange:
And probably if you're rich enough to own that many newspapers, you know what to do with those increases.
The second point is that we need a new bubble, and new technologies, especially if we have so much science fiction about this technology.
Since a lot has been written about ChatGPT and parrots, I would like to clarify in simple terms what exactly this technology is. Let's start from the beginning.
What is text data mining?
Let's use IBM's definition:
Ultimately, it means using a computer for the purpose of extracting knowledge from data (data mining), and specifically in a type of data which is, in fact, text.
How would you do without AI?
Well, I speak for things I've done. You would get yourself an infrastructure, let's say, Cloudera or Cloudera/Teradata, to hold huge amounts of text and to have the ability to access it in a reasonable amount of time. You would get the pipeline to add data, and given the data you should ask yourself how you are going to extract knowledge from the data.
You can do it programmatically: become an expert on, say, Spark and write your own jobs. Or use all job types that are allowed by Cloudera/Hadoop.
In this case, you must be a programmer, and know what we are talking about: if you want me to write you the carbonara recipe using biblical tones, you will first have to extract "the carbonara recipe", then learn "biblical tones" and then do some of product of the two.
If you're not that good at it, you can always use ready-made tools to build jobs, such as a Datameer. Using that, you can use mathematical tools without knowing what they are, and even deep learning tools without having expertise: clustering, decision trees, correlations, and others: Datameer is a visual tool that allows you to do things without necessarily understanding them. (what could possibly go wrong?)
Even to represent the results, you could use a tool, like Tableau, that makes it easy for you to represent the data in a humanly understandable format.
As you can see, the human being is at the end of the process, he expects a verified and mathematically treated datum, and the only problem is to do everything by understanding what language the computer speaks.
However, there is another possibility: if you are not good at programming and you are not even very good with IT tools, we could put something even easier in front of you.
In this case, all we will do is put in front of you an interface capable of understanding your language, which will listen to your requests arriving in natural language, and then return what you have asked for.
ChatGPT is a text data mining system, which uses an AI as an interface to understand your requests and make a job with it, and then goes back to using an AI to return the answer in natural language.
In fact, it contains a gigantic collection of texts, from which it does data mining, and returns something.
Is this the next electricity? Well, nowadays without electricity you can't live, without ChatGPT it's much easier to live.
Don't get me wrong: is it interesting? Very interesting. Is useful? Well, to many, who do concept work, it will be useful. Will it make jobs disappear? Well, if your job is to translate simple texts (I don't know, manuals) you could lose your job. (and maybe I'll understand what "if I want to boil my drill I have to push the rabbit with a clear storm" means, but maybe they translated it from Chinese using google.)
And so, very well: it's very interesting, it's useful to many, harmful to some, but if you ask me if it will be the next electricity, my answer is that for now I don't see it.
Why then are they attributing everything in the world to him, including the ability to make you rich on a hundred dollars?
Because as Cuccia said, the middle class is also the oxen park. We need to convince the oxen to hunt the lira and throw it in some sector, which sooner or later will have the usual story of failing banks, the money they keep as usual and the tears they cry… always the same.
The problem is that after plucking them with the Blockchain and Tinder for Gatti and with venture capital and startups and incubators, now we need to convince these fools that there is still time to make money with artificial intelligence.
Of course it will end as usual. We will hear about AI here and AI there, they will explain to us how many things we will do with AI, they will show us photos of young creatives with an artichoke-shaped desk and a room for lesbian soccer in the company, whoever will word AI will receive millions in venture capital somewhere, until the bubble bursts and as usual the middle class, sorry, the ox class, will lose the money they had invested.
And all of this is being born now, before your eyes. In your favorite newspaper.
Only thing I'm wondering is how many times the same boring chicken plucking trick is still going to work.