search burger
search ×

Is Google behind on artificial intelligence?

Yes, but (maybe) nothing serious


"Born later, born late, born wrong" is one of the refrains that Inter fans like me have been hearing all their lives repeated mercilessly by their Milan cousins on the occasion of their victories. In a (not so) hypothetical feud between big tech supporters this refrain could fit very well with Google's predicament in the artificial intelligence (AI) development challenge. That of finding itself behind and not yet able to close the gap with OpenAI and Microsoft.

With one difference: epithets among soccer fans hardly ever change over time, no matter what, while the race for AI is so frenetic that the scenario could change very quickly, along with observers' judgments.

The delay on AI

OpenAI's launch of Chat GPT in late 2022 shook Silicon Valley and seemed to take everyone by surprise, Google in the lead, which had been investing in AI long before that. The question immediately was not so much whether the Mountain View company was behind in developing such a generative product, but how far behind it was.

So much, judging from Google's first response: Bard, a chatbot whose demo was revealed in February last year still with obvious flaws. In fact, its unveiling was so disastrous that its stock lost 7 percent on the stock market. Since then Google has been trying to make up ground.

The arrival of Gemini

In December 2023 it released Gemini, a multimodal artificial intelligence that was born as an evolution of Bard and a version of which was made freely available to the public. Google's flagship AI model, Gemini Ultra, is by many parameters slightly better than Chat GPT-4, according to the company's tables.

Looking at the glass half full in about a year Google has equalized the performance of a tool that OpenAI has been working on for at least five years. Wanting to be pessimistic, Big G has made only minor improvements to something the Microsoft-funded startup completed work on at least 12 months earlier. And it will most likely be outdone again soon.

Gemini has so far not even been enough to boost Google's stock price. The lag is still affecting the market as the company has lost 100 billion dollars in capitalization since the beginning of the year, a drop of 6 percent.

What if instead

The challenge to govern AI is obviously hotly debated among engineers and tech enthusiasts on Reddit. And that's where I came across an interesting hot take: an opinion against the grain, perhaps a bit exaggerated, based on, however, some hard facts and reasonable predictions. As well as on a palatable metaphor.

"Just because a bunch of people arrived early to the party and had to prepare snacks doesn't mean the main course won't be great," the user writes. "Google has been slaving away in that kitchen longer than anyone else and has produced quality food in a wide range of cuisines".

Clear no? According to this opinion, one should not look too much at Gemini, because math is the key to artificial intelligence and so far Google is far ahead of the curve because it has been researching "real" learning for years.

For Google, there is hope

A view actually shared by several insiders and various nerds. I summarize it: Google is chasing, needs to review its priorities and perhaps choose a different business strategy, but it should not be given up for dead because it has the resources, a solid structure and some of the best minds around.

It probably just needs time. Like some teams to get to the top.



Illustration by Gloria Dozio - Acrimònia Studios