Jack Hidary, CEO and co-founder of Sandbox AQ, predicted that 2024 will become a milestone year for humanoid robots with Artificial Intelligence, on Thursday from the stage of the Greek House Davos 2024 conference. artificial intelligence and quantum technology company, which sprung from the bosom of “Alphabet”, the parent company of Google. “In the next 36 or 48 months, we’re going to start seeing embodied AI. And that’s a fancy way to describe robots” assessed Hindari, who even disclosed that he was invited to Greece for meetings with the prime minister and ministers.

But why haven’t the robots already taken off? “Because our world is meant for people. For example we have knobs and all kinds of devices and tools designed for people. So a robot with wheels and a clumsy little hand is not going to make it in this world. But in the last five years, in the human world, five different technologies have converged, to arrive at this moment in 2024. And I make a prediction here at “Greek House”: in 2024, we will see a “ChatGPT moment” for the humanoid robots. We will see demonstrations of real functional humanoid robots, so exciting, that will convince us” he said and added that Artificial Intelligence (AI) can offer important solutions to humanity, in two areas, for which all states – whether we are talking about the US whether for Greece or the UK – they are constantly struggling to solve problems: health and education

This does not mean, however, that the entry of robots into health or education will displace doctors, nurses or professors and teachers: “we want doctors and nurses to focus on their work,” he said, adding that robots can do tasks such as the disinfection of hospital rooms from bacteria and pathogens. “We can’t get rid of them without UV light, which can be dangerous for humans. A robot could be sent into a room, close the door and disinfect it so it’s ready for the next patient,” he noted, adding that AI could also help radiologists evaluate X-rays or doctors make diagnoses. “scanning” on behalf of voluminous patient files.

More female doctors, mentor teachers and a special school

Accordingly, in education, the use of Artificial Intelligence and Large Language Models (LLMs), but trained with the right data, could make a difference. “Yesterday, here in Davos, Sandbox AQ and others launched the world’s first LLM RAG. To explain what this means in practice, let’s look at a scenario. We want more women to become doctors. But where will they be trained? Are we going to magically find the money to build five new medical schools? This is not going to happen. What we can do though, is take the power of LLMs and create personalized education (…) The problem is that LLMs also draw data from science fiction novels or other entertainment genres (…) What we announced yesterday are large language models, trained on data specifically for training people who will become doctors and nurses,” he noted, disclosing that the TN model was given a corpus of 700 medical and nursing textbooks, licensed, validated and peer-reviewed, as part of what it is called “Retrieval Augmented Generator” (RAG- Retrieval Augmented Generator). The combination of LLM and RAG is the future of education” he emphasized, clarifying that with the use of such tools, professors will not be replaced, but will be able to do more creative work.

As an example, he mentioned the example of a school in the greater Austin area in the USA, the “Alpha School”. In this special school, during the first two hours of each day, from 9 to 11 in the morning, the children do not typically attend a lesson, but interact – using AI – with an iPad that offers material personalized to their needs and adapted to their own learning pace. “The teacher is there, but more as a coach and mentor, than as the person monologuing in the room. We’re not “getting rid” of teachers, but empowering them with highly personalized interaction tools, based on AI trained not on random stuff from the Internet, but on the right material,” he summarized.

Naive talk and AI as a tool in the hands of North Korea

We live in an age where everyone – citizens, organizations and governments – are talking about Artificial Intelligence. But are the “right” policy discussions taking place on IT issues? According to Hindari, although there are many policy discussions and they are necessary, yet as other forms of AI emerge, these discussions will appear naive. As he said, much of the current discussion on the subject revolves around AI-Terminator and a dystopian, post-apocalyptic future. That’s a conversation that needs to be had, of course, but one that doesn’t address the current threat: the fact that AI is for hackers around the world the tool of their dreams, he said. The same, he added, also applies to entire states, which invest large sums and employ thousands of people, in order to utilize IT not only for the defense of their country, but also for (cyber)attacks in other countries and especially in the companies that operate in them, with the aim of stealing intellectual property or hitting with ransomware-type malware.

Typical in this respect is the example of North Korea, in whose budget there is, according to Hidari, a relevant section. The country recorded $650 million in revenue last year from the use of ransomware. “In the hands of the North Koreans, artificial intelligence becomes a super tool for cyberattacks and phishing,” he said, adding that when we think about a national policy plan around AI, this is a very important parameter to consider.

The impact that the combination of AI and quantum technology can have was mentioned by Dr. Nadia Carlsten, Head of Product at SandboxAQ, who, however, referring to the maturity of quantum technology, said that we need to realize that the practical applications of quantum computing are far several more years. However, a significant advantage is created for businesses that try now to understand what quantum technology can – and cannot – do.

“In the pharmaceutical industry, for example, it’s good to know that nobody disputes that quantum computers will be better at computing molecular systems. That’s because molecules are quantum systems (…) So if you’re in this industry, you should do everything you can to learn more about quantum computers now so that you can one day leverage them as part of your toolchain. your. The other application that quantum is famous for is its ability to break RSA encryption. That hasn’t happened yet because quantum computers aren’t powerful enough to do it. But we know it will happen at some point. The question is when, and depending on who you ask, you’ll hear a different story. But what we’ve seen is that the companies that are preparing for that day are far ahead in terms of general security,” Carlsten said, adding that every industry will be affected by quantum, if only because of encryption: “the medicine, biotechnology, materials science, anything that has to do with creating new molecules, will be drastically affected by both quantum and TN and advanced simulation toolsets,” he explained.

Quantum ‘GPS’ and how many qubits*… make a difference

He added that although we usually talk about quantum computers, there are other interesting quantum technologies, such as quantum sensing, which is the use of quantum systems to make measurements. “One of the fun things we can measure is magnetic fields. Earth has a magnetic field. The human body has a magnetic field. So there are two applications, one is for medical imaging (through the body’s magnetic field) to make diagnoses (which is already being tested in some hospitals) and the other is navigation. Using the same technique, i.e. receiving magnetic fields, we can actually calculate a location, without having to use GPS (…) Today we see this capability as a safer complement to GPS, which cannot be interfered with, as the signal is the Earth’s magnetic field itself. In the end, who knows, maybe one day it can replace it” he emphasized.

When might we see the first practical applications of quantum computers? “Right now there is a race to achieve what are called ‘logical qubits’, that is, multiple physical qubits that actually work together as a unit, a kind of computing unit. When we get to more than a handful of “logic qubits,” that’s when interesting things start to happen. QuEra, a quantum computing company, has just announced a (quantum processor with) 48 qubits. That’s a big milestone, but it’s not enough. At the current error rates, we’ll need hundreds, thousands of qubits. So it will take time (…) I think what we will see (when we achieve this) are problems of increasing importance that will start to be solved. We will be able to do something quite interesting in the region of 1000 logical qubits. And then it would take 100,000 to simulate a molecule that is useful for drug discovery.”_

*The quantum bit, or more commonly qubit, is the elementary unit of quantum information.