Ouch, ouch, ai! Did the Hype admit? Let’s see a little. The guardian of August 25 has cooled the enthusiasm with a title that is worth more than a report by McKinsey: “Is the Ai Boom Finally Starting to Slow Down?”. Inside there is the study of the MIT which clearly says that 95% of the company projects in AI has not produced concrete benefits, namely that the great enthusiasm of the last two years translates for now in a huge waste of budget and slides (with the tiktok paradox that uses Ai not to free creativity, rather to fire human moderators, oh well, you know how I think of Tik Tok).
However, the picture is not so clear, things are always complex, the complex things even more complex. Mit’s same report tells the so -called “Shadow Ai Economy”: while official projects fail, more than 90% of the use of chatgpt employees or Claude Sottobanco to work better, even if the company does not know. An important detail? Maybe that maybe maybe not. So much so that Raffaele Gaito, a very active popularizer on YouTube, takes the data to his advantage (being a business consultant): if the employees use it, then it means that it works, and therefore 95% of failures should be reduced.
Mmm … does not make a turn, if the purpose is to bend the data to Gaito’s work. Too bad that using does not mean integrating (and he knows it very well too). Writing an email with chatgpt or being summarized a pdf or generating a Linkedin title is not transforming productivity, it is like having a digital intern for free (and that in any case you have to check, otherwise in any company you risk making small and large mess). However, I understand that if your job is to sell the revolution as already in progress, even when the numbers say the opposite, there is.
In the meantime Ennk with his “Ai without hype” does not discount, also because as a researcher he has no interests if not to read quintals of paper and make them digestible for experts and less experienced (with unmissable videos, such as his very popular Technicismi column). It reiterates that the numbers of the MIT are very credible, because they describe what you see: expensive projects that capture the imagination more than the budget and the real use that still remains very fragmented (much much more than they seem to those who believe that today everything is to when on the contrary are emerging many critical issues).
Among other things, the market adds another note on a margin, of the opposite sign: the last article by Reuters, entitled “Nvidia Ceo Says ai boom Far Over”, confirms that the revolution is not set at all. Jensen Huang has relaunched: 3-4 trillions of dollars of investments in infrastructure at 2030, with the hyperscalers ready to spend. It is proof that the trajectory is there, but it is slower and more expensive than the consultants and conference stages and those who have to make videos on YouTube to convince companies would like to make believe.
Without hype, as a curious writer and external observer, I have no doubts: the Ai will really change the processes and markets, but not with the speed that we are telling each other, as basically happened with the Internet. It is not that Google or Amazon were born together with the web, who already existed in the seventies (Remember Arpanet?), In 1983 he was born the Internet with the TCP/IP protocol (when we children at the most programmed played games in Basic, and there was nothing really online), Amazon arrives in 1994 (as a bookshop online and bezos bet), Google in 1998, and only in the two thousand years. We begin to seriously speak of the web economy, e-commerce, social networks, digital advertising and so on, up to Facebook (2004, now the social of the old) and YouTube (in 2005) and we had not yet seen the first iPhone. How long huh? (Relatively speaking, in the arc of human technological progress is a flash, goes without saying, but the same human species has been there very little, two hundred thousand years out of four billion years of life on this planet, in the scale of life “man is only the paint of the tip that covers the tip of the Eiffel tower”, to use a metaphor of Mark Twain that makes the idea very much).
The same thing will be worth the AI, with its accelerations and decelerations, independent of the enthusiastic or not. When it stops being a media circus and becomes an invisible part of the mechanisms, we can talk about revolution, and when the revolution will arrive, it will arrive without announcements (the same Sam Altman has learned the danger of bombastic ads), because there is never a precise day in which someone says “today there is the revolution” (except in the political revolutions, which always end badly): we will realize that it is already there, which was already there, look elsewhere.
And if everything goes wrong, as the usual un integrated apocalyptic apocalyptic claim, it will be like in the quotation of the talking heads that Breat Easton Ellis put at the beginning of American Psycho: “And when everything went to Catafascio, nobody had a lot of mind”.