Over the past 12 months, we have seen significant advancements in various technology areas, ranging from electric vehicles has mixed reality technologiesbut much of the conversation was dominated by artificial intelligence (AI).
As big language models – the current gold standard, based on the neural networks that power everything from Windows Copilot to ChatGPT – have gradually improved in 2024, this is the year when the existential risks of language AI have become disturbingly clear.
Another area poised for radical transformation is quantum computingwhere new progress was reported every month. Not only are machines getting bigger and more powerful, but they are also becoming more reliable, as scientists get closer to machines that outperform machines. best supercomputers. Some of the most important advances have taken place in the area of error correction, a key problem that must be solved before quantum computers can realize their potential.
And in the world of electronics, scientists have moved closer to realizing a hypothetical component known as “universal memory,” which, if realized, will transform the devices we use every day.
Here are the most transformative technology developments of 2024.
We are closer to understanding the existential risks of AI
This year, AI companies have released increasingly improved extensive language models, including OpenAI’s o1, the Evo model for predicting genetic mutations and the ESM3 protein sequencing model. We also saw better methods of training and processing AI, such as a new tool that speeds up image generation up to eight times and an algorithm capable of compressing these models so that they are small enough to run locally on your smartphone.
But it was also the year when the existential threats associated with AI became more visible. In January, a study showed that widely used security training methods failed to suppress malicious behavior in models that had been “poisoned” or designed to display harmful or undesirable tendencies.
The study, described by its authors as “legitimately frightening”, found that in one case, a malicious AI had learned to recognize the trigger for its malicious actions and thus attempted to hide its antisocial behavior from its human handlers. Of course, they could see what the AI was actually “thinking” at any time, but that wouldn’t always be the case in the real world.
We are opening a viable path to useful quantum computers
It’s been a busy 12 months quantum computing research. In January, quantum computing company QuEra created a new machine with 256 physical qubits and 10 “logical qubits” — collections of physical qubits linked together via quantum entanglement – which reduces errors by storing the same data in different locations. At the time, it was the first machine with built-in quantum error correction. But teams around the world are trying to reduce the error rate on qubits.
The flagship development in error correction was revealed in December, when Google scientists announced that they had built a new generation of quantum processing units (QPU) which have taken an important step in error correction, where as you increase the number of qubits, you correct more errors than you introduce. This will result in an exponential reduction in errors as the number of entangled qubits increases.
THE new 105-qubit Willow chipwhich succeeds Sycamore, managed to obtain a breathtaking result in terms of comparative analysis, by solving in five minutes a problem that a supercomputer would have taken 10 seven billion years to solve – a quadrillion times the age of the universe.
‘Universal memory’ is getting closer to reality: here’s what it means for the devices we use
Although this year brought several innovative computer components, including a new type of data storage that can withstand extreme heatas well as a DNA-infused computer chip – some of the most important advances have taken place in the development of “universal memory”. This is a type of component that will significantly increase computing speed and reduce power consumption.
All computers use two types of memory at once: short-term memory, such as random access memory (RAM), and long-term storage, such as solid-state drives (SSD) or flash memory. RAM is incredibly fast but requires constant power; All memory stored in RAM is deleted as soon as a computer is turned off. SSDs, on the other hand, are relatively slow but can retain information without power.
Universal memory is a third type of memory that combines the best of the first two types – and, in 2024, scientists have moved closer to realizing this technology.
Earlier this year, scientists showed that a new material called “GST467” was a viable candidate for phase change memory – a type of memory that creates 1s and 0s of computer data when it switches between high and low resistance states in a glass-like material. When it crystallizes, it represents 1 and releases a large amount of energy. When it melts, it represents 0 and absorbs the same amount of energy. In testing, this material was found to be faster and more efficient than other universal memory candidates, such as ULTRARAMthe current leading candidate.
Other candidates are also promising – and bizarre. In April, for example, scientists proposed that a strange magnetic quasiparticle known as “skyrmion” could one day be used in universal memory in place of electrons. In the new study, they accelerated the skyrmions from their normal speed of 100 meters per second (about 225 mph or 362 km/h) – which is too slow to be used in computer memory – to 2,000 mph (3,200 km/h). .
Then, towards the end of the year, scientists accidentally discovered another material that could be used for phase change memory. This has reduced the energy requirements for data storage by up to a billion times. This discovery happened entirely by chance, demonstrating that in the world of science and technology, you never know how close you are to a major breakthrough.