Category Archives: Research

Artificial Intelligence, Research, Science, Society, Technology

Superintelligent AI and its threat to humanity

ARTIFICIAL INTELLIGENCE

Intro: Humanity faces an uncertain fate as experts brace for superintelligent AI. The tech industry claims looming “singularity” will change everything

Every time one of the world’s top artificial intelligence companies unveils a new system, employees at the US research organisation METR put it through its paces. Its ability is tested to complete a series of increasingly complex tasks.

The tasks are measured by how long each one would take a skilled human. They range from trivial arithmetic (two seconds) and completing a game of Wordle (13 minutes) to building complex military satellite software (taking a human expert 14.5 hours).

The test then serves as a gauge as to how capable AI has become – and where it might go.

The first version of ChatGPT, released in 2022, could only perform simple tasks that would take a human a few seconds.

But as AI systems have become more powerful, they are able to complete more complex actions that would take humans hours or days, such as breaking into a medical website and downloading all its data.

METR has found that AI capabilities are doubling in power every 196 days. Plotted on a graph, this progress starts slowly then rapidly accelerates to a near-vertical plane.

Converse with anyone in the AI industry for any length of time and the likelihood of them pulling up a version of the chart approaches 100pc, to the point where it has become a meme in its own right. It is being referred to as the most important chart in the world. The chart goes off the scale.

Last month, the AI lab Anthropic announced it had developed a new system, called Mythos, that it said was too powerful to release to the public because of its ability to find gaping holes in online security systems.

When METR’s researchers released the results of Mythos’s capability and function, they scored the system at 16 hours – meaning the world’s most powerful AI can now automate tasks that would take a human two full eight-hour shifts.

Nonetheless, they said the model was “at the upper end” of their ability to test. In other words, progress has become too fast for them to measure.

Not everybody is convinced by the results because the test only measures if a machine can do something half the time, not if it can do it consistently. The METR chart has, however, captured many people’s imaginations for two reasons.

First, the exponential growth looks strikingly similar to “Moore’s Law”, the maxim that has governed the electronics industry for more than half a century, stating that microchips roughly double in power every two years.

Second, it measures abilities, rather than intelligence. While many AI “benchmarks” resemble university exams and gradings, dealing in abstract reasoning or maths, the METR test studies whether AI can actually work.

It suggests that on current trends, vast amounts of human tasks could be automated in the next couple of years – including, most crucially of all, the art of developing AI models itself.

At that threshold, known in the tech industry as “recursive self-improvement”, all bets are off.

The concept is closely linked to superhuman AI because an AI that can make itself smarter could act like an evolutionary chain reaction, rapidly building to a system vastly more capable than mankind.

AI would have become – as IJ Good, the Bletchley Park codebreaker, predicted in 1965 – “the last invention that man need make”. Almost Orwellian in thought.

For 60 years, the idea seemed out of reach. But much of Silicon Valley believes this is about to change – and the US government is starting to notice.

The vast majority of people’s experience of AI has not changed much in the last couple of years. The release of ChatGPT in 2022 generated an initial flurry of excitement and fear in equal measure but, since then, progress has been less obvious.

The AI experience for many people comes in seeing an obviously fake video on their social media feeds, seeing an AI overview at the top of their search results, or having a bot that “helpfully” offers to summarise their emails.

But at the coalface, people are rapidly bringing forward their timelines for the day that superintelligence arrives.

Pages: 1 2 3 4
Standard
Health, Medical, Research, Science

Blood test that can detect Alzheimer’s 15 years before onset

ALZHEIMER’S DISEASE

A SIMPLE blood test can detect Alzheimer’s disease up to 15 years before symptoms begin, a major trial has found. It paves the way for a national screening programme.

The trial found that the test was as accurate as the current gold standard for diagnosing the condition.

For the first time, doctors were able to say if a person had a high, medium, or low chance, of having the disease – ruling out further invasive procedures.

Experts have said it would “revolutionise” diagnosis, making Alzheimer’s as easy to test and detect as for other routine health conditions such as high cholesterol.

Patients could expect results within days of visiting their GP, rather than the years it currently takes to get a diagnosis. This could have huge implications for future treatments, removing the barriers for a diagnosis – such as long waits for spinal taps or brain scans – and speeding up trials.

It could also pave the way for screening over-50s once more effective treatments become available.

Made by diagnostics company ALZpath, it was found to be 97 per cent accurate at detecting traces of the “tau” protein, which was linked to developing Alzheimer’s disease during the eight-year trials. These proteins start to build up on the brain 10 to 15 years before symptoms start showing.

Researchers in Sweden found high levels of the “tau” protein in the blood test corresponded to high levels of Alzheimer markers seen in expensive diagnostic brain scans and painful lumbar punctures.

The more of this leaked “tau” brain protein in the blood, the more likely or advanced the Alzheimer’s disease was in the tests involving 786 people. Growing evidence suggests biomarker changes like these can be detected in the blood years before other signs of the disease appear in the brain.

It means if scientists can find a way to stop these protein levels from rising, they could effectively halt Alzheimer’s in its tracks.

With breakthrough treatments such as donanemab and lecanemab on the horizon, experts say it is vital to have quick and reliable diagnoses. Professor David Curtis of University College London Genetics Institute said this was “one half of the solution”, while we await effective treatments.

He added: “This potentially could have huge implications. Everybody over 50 could be routinely screened every few years, in much the same way as they are now screened for high cholesterol.”

Around 900,000 people in the UK live with dementia – with Alzheimer’s the most common form. The growing ageing population means numbers are expected to rise to 1.6million by 2040, making a cheap screening tool vital to get to grips with the challenge.

Alzheimer’s Research UK analysis found 74,261 people died from dementia in 2022 compared with 69,178 a year earlier, making it the country’s biggest killer. While previous blood tests have shown promise, these findings have caused particular excitement given the high accuracy levels, large study size, and because the test already exists commercially.

It is also the first time a blood test has been found to be at least as good as a painful lumbar puncture or spinal tap for detecting elevated levels of the tau protein, according to the research team at the University of Gothenburg, Sweden.

Lumbar punctures involve taking fluid from the patient’s spinal cord. The inexpensive tests – priced at around £150 – could also be used to monitor a patient’s condition, allowing more tailored trials or treatment in future.

Dr Richard Oakley, of the Alzheimer’s Society, urged that more research would be needed, but said: “This study is a huge welcome step in the right direction as it shows that blood tests can be just as accurate as more invasive and expensive tests.

“It suggests results from these tests could be clear enough to not require follow-up investigations for some people living with Alzheimer’s disease, which could speed up diagnosis.”

The tests would need regulatory approval before widespread use. But they could form part of NHS trials starting imminently and looking to roll out blood tests for Alzheimer’s within the next five years.

The scientists’ findings were first published in JAMA Neurology.

Standard
Research, Science

Scientists reveal that the proton’s size is moveable

PARTICLE PHYSICS

Intro: The radius of the proton, a subatomic particle, seems to vary depending on how you look at it

THE proton, one of the building blocks for all matter, is proving to be an awkward customer to size up. If you look at its charge, it will have one radius, but if you look at its mass, you will see a different, smaller radius.

A new picture of the proton is emerging. In the 1960s, experiments that fired electrons at protons revealed that the latter contained point-like, electrically charged particles that we now call quarks. A proton has two up quarks and a down one. Quarks were later found to be bound together by particles called gluons.

We now know more about quarks and how far their electric field extends in space, which is sometimes called the radius of the proton. But we know less about gluons, which contain most of the mass of the proton in the form of energy, because they are chargeless, and so much harder to investigate. Seeing how they are distributed can tell us about how the proton’s mass is arranged and its structure.

Scientists at the Argonne National Laboratory in Illinois have probed the proton’s gluons with particles called J/psi mesons. This is possible because even though gluons don’t have electric charge, they have a property called colour charge, which comes from the strong nuclear force, one of the universe’s four fundamental forces. J/psi mesons are made up of a charm quark and its antiquark, which also have colour charge and so are capable of interacting with gluons.

The researchers fired a beam of photons at liquid hydrogen, which is comprised mainly of just protons, and the photons interacted with the protons. These collisions produced short-lived J/psi mesons. By measuring how many of these were produced, the research team could calculate the proton’s mass distribution using quantum mechanical models that describe gluon-quark interactions.

Their results suggest that the gluons’ mass is confined to a dense core in the proton’s centre, while the charge from the quarks extends to a second, larger radius.

They also compared their results with predictions from another model of the proton, which agreed in some places and diverged at others, suggesting that the new figures need validating with more precise experiments or one that probe proton structure in a different way.

If it is confirmed, it will be a very interesting finding because it tells us something quite deep about how the proton’s constituents behave from a spatial point of view.

A different internal structure could have implications for calculating other proton properties, such as spin, angular momentum and energy distribution, which many sensitive experiments rely on. But some of the new proton findings rest on models used to calculate them, which haven’t proved entirely reliable in the past.

The results follow another revelation about the proton’s internal structure. Last year, a research team found that the proton can contain a much heavier charm quark, in addition to the three regular quarks, but asked: ‘Does the mass radius become larger or smaller?’

. Further understanding on quarks can be found:

Science Book: Physics

Standard