NEWSLETTER

Edition 105

Daniel McKinnon

1. One thing that helps

Factory of the Future

It’s a supersized printer, but it’s more than that. The University of Maine has developed a huge 3D printer that can be an all-in-one solution to manufacturing: Factory of the Future 1.0.

Hailed as the largest thermoplastic-polymer printer in the world, it can print a 600-square-foot house in less than four days. Its ability to perform additive manufacturing (3D printing), subtractive manufacturing (cutting or milling materials), and its robotic arm operations make it unique. This allows Factory of the Future to build a structure and install wiring and plumbing systems simultaneously. Sensors and artificial intelligence allow the printer to make adjustments to ensure smooth manufacturing.

Factory of the Future will be primarily used to print houses made of recyclable materials. This reduces the cost of ownership along with reduced carbon footprints. The tech can also drastically reduce the time and resources needed to manufacture vessels like boats and flying machines. I hope it doesn’t break as easily as those all-in-one printers we use at home.

2. One to be wary of

AI Tricksters

Is AI tricking us, or are they too good at finding shortcuts?

AI has been proven to be “deceptive”. Meta’s AIs, Cicero and Pluribus, are particularly good examples. Cicero was trained to be honest and never backstab in playing the military strategy game Diplomacy. However, in a new paper, researchers claimed that, unlike its namesake, Cicero was anything but honest, deceiving other players and breaking deals. Pluribus, an AI poker player, was so effective in bluffing that releasing it could wreak chaos on the poker community. Even systems like ChatGPT have been known to use deception, like pretending to be vision-impaired, to solicit help in solving Captchas! It even engaged in insider trading (during a test).

Before we cry, “AI is so evil”, we must remember that these are AI systems, not people. What we call deception may be AI doing its best to achieve its goals unexpectedly. This is a consequence of the black box problem in AI: we cannot know how and why they act the way they do. AI will be deceptive. So, before using them in the real world, we should weigh their benefits versus their potential harm.

3. One to amaze

Journey to a Black Hole

Remember this scene from Interstellar?

If you want to immerse yourself more, NASA has released two videos that simulate travelling into or around a black hole.

Let’s take a trip around a black hole.

As you approach the supermassive black hole, you see it surrounded by a hot glowing disk of gas. Speeding closer to the black hole, you see the sky distorted and mirrored. If the spacecraft successfully passes the event horizon, the black hole appears to split. But what happens if you fall into a black hole?

Given the choice, you’ll want to fall into a supermassive black hole 🙂 . It’s because larger black holes tend to have their event horizons farther from their singularity. That gives you a chance not to get immediately spaghettified – stretched thin due to tidal forces. Approaching the photon ring, we can still see light sources, albeit the ring itself appears distorted. Crossing the event horizon, your image will be frozen for outside observers to see. Beyond that, you get spaghettified plunging into the singularity, but of course, you’ll die before you even reach it.

Mesmerising as these simulations can be, one can only be amazed by the calculations that went into producing these two videos. A conventional laptop would take around 10 years to do these calculations, but NASA used the supercomputer Discover to render these in five days. Five days and 10 terabytes of data!

A 2-minute future guide, direct to you, once per fortnight.
Subscribed Successfully!
Oops! Something went wrong while submitting the form.