It was such an insignificant project that even most of the employees did not know about it and the supervisory board was not informed about it. When OpenAI launched ChatGPT three years ago, in November 2022, it was a prototype. A preview of what big language models can already do. So far, only developers and enthusiasts have played with them.

The name made little sense to anyone outside the developer community.He was referring to the abbreviation GPT, generative pre-trained transformer. The original name was even more arduous: Chat with GPT. “We’re much better at R&D then we are at naming services,” he glossed Sam Altman, then and current head of OpenAI. it didn’t matter much before. No OpenAI project had gained a mass audience until then. Their image generator, which was supposed to do this, was not open to the public.

But ChatGPT – despite the name, which people still mumble around to this day – has captivated users all over the world with amazing speed. Within two months,it gained over a hundred million users and became the fastest growing service ever. Soon, studies began to come out that estimated that artificial intelligence could soon drastically transform the job market.Three years later, it is indeed clear that change is slower. Especially as deploying AI is far from just a technical issue. The structures in businesses and corporations are – fortunately – tailored for human workers and have a lot of inertia. For now, employees in most companies with chatbots are experimenting rather at their own risk (and often at their own risk). So far, the impacts of artificial intelligence are mainly felt by external experts and entrepreneurs. graphic designers, programmers, translators, copywriters. But that’s just the begining.At least according to the visions of those who develop and sell AI.

AI salvation or AI destruction? What will the future bring?

How do optimists and skeptics see the future with artificial intelligence?

We will gradually explore six visions of how artificial intelligence (AI), general artificial intelligence (AGI) or even superhuman superintelligence (BUT) will t”Our main goal at OpenAI is to ensure that general artificial intelligence – when it becomes a reality – serves all of humanity,” he emphasized.

## Gentle singularity according to Altman

Since 2019, the OpenAI company has gone through a number of changes. She received a financial injection from Microsoft, which allowed her to start training orders of magnitude larger and smarter models. It started the chatbot mania and provoked even the hitherto sleeping giants into action in the field of generative AI. OpenAI’s supervisory board removed Sam Altman from the position of chief, only to reinstate him after pressure from Microsoft and employees. But the enthusiasm and obsession with general artificial intelligence has not abated. Maybe just the opposite. With each new and smarter LLM (large language model), the debate about when “this will be it” intensifies.

Altman also worked on negotiation skills. He still speaks informally and directly, but in recent years there is more balance behind his answers. Just as his chatbot learned to think first and then write a response, Altman seems to be increasingly aware that he and other AI company bosses suddenly have more in their hands than just the fate of their companies.

As a small laboratory, OpenAI could copy texts from the Internet, train its models on them, and no one really cared. But suddenly experimental models became globally available tools that fundamentally change entire fields of human activity. OpenAI is facing several lawsuits alleging that their chatbot’s responses encouraged young people to commit suicide and induced psychosis-like states in others. Teachers around the world are looking for ways to incorporate chatbots into education without falling into the trap of lazily copying texts here and there.

Altman is aware of the seriousness of the implications. Or at least he realizes he should act like he doesn’t care. Just as we can’t see inside the AI’s head, we don’t know exactly what plans are going through the mind of the person who set off this avalanche.After all, people who have worked with Altman often complain that he is often selective in what he says to whom. He was said to be one of reasonswhy the board fired him: he was manipulative and sometimes outright lied.From what Altman says and writesbut we can build a picture of the utopian vision towards which it is moving in the long term. He calls it the “gentle singularity” – alluding to the point of no return.

An army of geniuses in the data center

Dario amodei, Altman’s former colleague, who together with his sister Danila decided to leave OpenAI in 2021 and founded the rival laboratory Anthropic, also focuses on scientific progress in his vision. They managed to win over some developers with their even stronger emphasis on security. Amodei, a trained biophysicist, also had to learn how to appear in the media. In four years, Anthropic has grown into one of the most crucial AI companies in the world, with investments from Amazon and Google.

“Intelligence is really powerful, but it’s not magic,” reminds amodei. He is therefore skeptical of predictions that AI could accelerate scientific knowledge at an exponentially increasing rate. At some point, they will run into physical limits, whether they are related to the speed of building hardware or the speed of biological experiments. But simultaneously occurring, he believes that there are a number of problems that “better-than-human intelligence” could solve relatively quickly (Amodei usually avoids the concept of superintelligence).

He hopes, however, that advanced systems – which he describes as “a whole country of brilliant scientists packed into a data center” – would fundamentally advance scientific knowledge.

“`html

AI and the Future of Work: Adapting to a Changing Landscape