This week on Talk Nerdy to Me, we examine spiders in space, Earth's unique habitability, and how AI can be a friend or a foe.

Staying alive, planet style

It’s alarming to think of Earth’s habitability as a precarious balancing act; however, that is exactly how researchers at the University of Southampton describe it. While new tools have steadily discovered more planets, the question remains: How likely is it for any planet to sustain life for long?

According to these researchers, doing so is no easy feat. A large number of events can render a planet lifeless, from the terrestrial (e.g., super volcanoes) to the extraterrestrial (e.g. solar flares). Given this, Earth’s long reign — approximately three to four billion years — as a habitable planet is extraordinary.

Is Earth’s sustained life thanks to some inherent ability of its to remain stable? Or is it a matter of luck? The researchers’ project was inspired to answer these questions and better understand the nature of planetary stability.

To do so, they simulated 100,000 randomly different planets one hundred times each, using the university’s supercomputers. Over the course of three-billion-year simulations, random climate-altering events were introduced.

Many of the planets failed most to all of their simulations —  only nine percent remained habitable at least once. The best performers were successful fewer than ten times out of one hundred.

AI foe

An international team of scientists set out to determine if humans are capable of controlling a super-intelligent AI. And, according to their theoretical calculations, we are not — it is an impossible computing feat.

The scientists determined this using a conceptual algorithm, built to prevent AI from harming people. The containment algorithm would simulate the AI’s behavior, assess its harm, and stop the AI if need be. However, in doing so, the algorithm could also halt its own functioning.

We would not know when this happened, and so we would not be aware if the algorithm had stopped analyzing potential threats. This inherent flaw makes such an algorithm useless.

The researchers also explain that we might not even know when AI becomes more intelligent than us, because — in a conundrum similar to their containment problem – human intelligence might not be able to detect such a development.

AI friends

We may be afraid of AI on a conceptual level, but this study shows that our interactions with them are surprisingly typical on a social level.

Researches at Tampere University measured the physiological reactions humans have when face-to-face with social robots. The human participants demonstrated heightened attention and emotional responses when the robots made eye contact with them. 

The results were surprisingly shocking to the researchers, because they had found in a previous study that eye contact only solicits physiological responses when participants know that they are being seen (e.g., the difference between having your camera on or off when on a video call).

Social cues, such as making eye contact, have evolved over time. They influence the ways we interact, the impressions we form, and the feelings we have. Apparently, this is true regardless of whether or not we believe the social signals are coming from a robot or a human.

Closing in on quantum computers

A milestone discovery made by researchers — in a collaboration between CEA-Leti and the Niels Bohr Institute — shows that everyday transistors can be used as qubits in quantum computers.

The regular computers we use today make a surprising number of errors, but they are systematically detected and corrected. Every computing calculation is carried out by multiple transistors, so that if one transistor makes an error, it is detectable to the computer. Scientists have been searching for a correctional measure such as this that would work with quantum computers.

And, as it turns out, the two-dimensional arrays these scientists achieved with everyday transistors offer a solution. If enough qubits are combined in a 2D array, they can similarly expose the mistakes of one another.

By producing these results with everyday transistors, the field can achieve previously unknown scalability and qubit production. This will be a huge step toward realizing science’s long-held dream of efficient quantum computers.

Loose spiders and loose ends

Through missteps, accidents, and oversights, experiments can turn into truly twisted tales. In this case, a specimen escaped, insects ran amuck, and havoc was wreaked.

The objective of this study was simple: to observe how spiders build their webs when not affected by gravity. On Earth, spiders typically build asymmetrical webs with their centers closer to the top. While sitting on their webs, spiders orient their bodies downward, because they can move more quickly toward their prey with gravity’s help.

So the researchers wondered: Would spiders build symmetrical webs in zero gravity?

In an article published by Science Daily last month, we see how the experiment veered off course so thoroughly that the results were never formally published.

Two spiders were selected – one as a backup – and shipped off to the International Space Station. At the ISS, the backup spider escaped its holding chamber and joined the primary spider in its observation chamber. Here, they built, deconstructed, and tangled their webs.

Meanwhile, their food supply, flies, reproduced at unexpected rates, overflowing the food bin with larvae. After one month, the larvae overgrew the entire observation window.

The data had become so muddled that, at the experiment’s end, the scientists were left with no scientific conclusions.

Fortunately, the project was relaunched three years later. Although it once again met unexpected challenges, the final results were not obscured.

There werre 14,500 images taken of 100 webs . In them, there were two decisive patterns: the webs built in zero gravity were more symmetrical than those spun on Earth, except when the lights were on. When on, the space webs were similarly asymmetrical to Earth webs.

In this way, the researchers’ long and twisting tale reached a conclusion and yielded a surprising insight: spiders have a backup orientation system. In the absence of gravity, they can orient themselves by light.