He doesn’t know why. When asked recently by journalist Connie Loizos, to TechCrunchOn why the general public has recently become so enamored with the capabilities of ChatGPT, an AI text generator modeled after GPT-3, Sam Altman, co-founder and CEO of OpenAI, the company behind the technology, has a skeptical answer. We do .
“Honestly, I don’t know, and I’ve thought about it a lot. We’ve been developing the model that ChatGPT is based on for a while, and I always had a feeling someone would build something like this. of achieving user experience. I always thought that the only way people wanted to interact with these models was through dialogue,” explains the CEO.
The answer is part of a longer interview in which Altman tries to shed light on some of the issues troubling the general public in the aftermath of the AI explosion.
OpenAI begins testing ChatGPT Professional, a paid version of its hugely successful AI text generator
These are summarized in one: And now he? Once something like ChatGPT, an AI capable of generating more or less concrete text with only a few slogans, has generated enough interest to make the weeks collapse, what can be expected from the rest of the AI tools? What are OpenAI plans? What about other big tech companies?
Finally, most important: How will it change people’s lives?
The OpenAI roadmap is more or less clear in Altman’s eyes. Beyond Microsoft investing $10,000 million in the company to integrate AI solutions into some of its products (for example, there is talk of a future Bing with these types of integrated solutions), OpenAI’s immediate future lies in the financial Goes through the muscle.
The reason is that, like all highly disruptive technology, AI is a very demanding industry in terms of investment. This means that, in order for the bet to be successful, it first requires a significant amount of money, so that the technology can be improved.
OpenAI knows this better than anyone, which is why the $10,000 million that Microsoft is willing to pay doesn’t know much about it. company, as you remember business InsiderIs in talks for a capital raise that could take it above $30,000 million in market capitalization.
In this way, Altman confirmed Development of AI capable of making videos, It will be added to the current functions already launched by OpenAI: image (Dall-E) and text (GPT-3), although he was cautious when giving a release date: “It may be soon, it’s a matter of time for us.” research project” is being developed again, although this too may take some time.
He also didn’t want to give a date for the launch of GPT-4, the development of GPT-3’s definitive leap toward what many know as Artificial General Intelligence (AGI). expected to apply.
ChatGPT and generative AI seem to be the next tech boom, but it may turn out to be a new bubble
It is a paradigm whereby machines overcome the current trial and error processes on which they currently base their operations. develop human intelligence while doing creative work.
Altman only said of GPT-4, “We’ll release it when we’re sure we can do it safely.”
“We don’t have real AGI, and I think that’s exactly what is expected of us and GPT-4. In that sense, it’s disappointing. The closer we get to AGI, the more difficult it is for me to answer Will happen.” The question of how it will be. Because I think it’s going to be a much blurrier, more gradual transition than people realize,” Altman developed.
A new society for a new technology
Opz’s CEO, on the other hand, was less elusive when it came to pondering the profound changes that many already predicted the AI revolution would bring.
This is already causing concern in areas such as education, which sees, for example, automatic text generation tools as a huge open door for students to submit their own writing as developed by robots. , who have barely received some instructions.
ChatGPT can be used for good, but like many other AI models, it is riddled with racist and discriminatory biases.
This sense of danger has led to cities such as New York, for example, restricting access to ChatGPT in school environments, although nothing can be prevented, the promoters of the measure themselves admit that students can use this AI from home if they wish. You can use.
In this regard, Altman revealed that OpenAI is experimenting, for example, with watermarking techniques to try to help teachers distinguish text written by AI from text written by a student.
However, he also issued a caveat: focusing on these kinds of solutions is a gateway: “There may be ways to help teachers explore GPT a little more, but honestly, a determined person can avoid these.” Sanctions are going to do it, and I don’t think it’s something that society can or should rely on in the long run.”
Instead, Altman proposes a change in mindset: “Calculators have replaced what we test in math classes, and Google has reduced the need to memorize data. reading or learning They are a more extreme version”, he recalled, adding that there would be benefits as well.
Speaking of benefits and drawbacks, Altman ultimately addressed an important question: How do you envision the best- and worst-case scenarios in a world dominated by AI.
With Bing and ChatGPT, Google will compete in search for the first time in 20 years
“The best case scenario is so incredibly good that it’s hard for me to even imagine it. I can hardly imagine what the world will be like when we go ahead with the systems we’ve developed so far. The good case scenario is so incredibly good That you sound like a lunatic talking about it.” He explained, on the one hand.
On the other hand, the worst-case scenario is simply the end: “The lights go out for all of us. It’s not that AI will one day rise up and decide to turn evil, but that I’m more concerned about the inappropriate use of technology.” “.