Welcome to the Robot Remix, where we summarise the week's need-to-know robotics and automation news.

In today's email -

• Automated sculpting
• Don’t steal IP from Apple
• EV factory tour
• AI drawing robots
• Please don’t build this knife throwing robot

Snippets

‌                                                              

Robot rock art - A robot sculptor has recreated the famous Selene Stallion that adorned the Parthenon in Athens 2,500 years ago. The system used scans of the original head and a heavy-duty robot arm to produce the statue in less than 24 hours. Why was the system was developed? The Greek Government is hoping to convince the British Museum to return their ‘borrowed’ marbles in exchange for the replicas.

A few bad Apples - An ex-Apple Car engineer has pled guilty to stealing self-driving secrets.  Xiaolang Zhang admitted to transferring 24GB of “highly problematic” data to his wife’s laptop and taking circuit boards while working on the self-driving car project from 2015-2018. Following his stint at Apple, he went on to work at Guangzhou Xiaopeng Motor Technologies, a Chinese EV startup... An honest mistake?

Faraday Future’s Factory - Check out this video tour of Faraday Future's new production facility. The electric vehicle manufacturer has allowed us to peek inside their Californian facility. The factory is slated to build Faraday’s flagship EV model, the TechLuxury sedan and has the ability to assemble 10,000 vehicles when operating at full capacity. Quite a few robots in there.

ABB joins Berkshire Grey’s Technology Alliance Program - ABB is looking to enhance its offering by providing more customers with AI-enabled robotic solutions. Check out next week’s Big Idea on Deep learning in robotics.

The dark side of technology - Racist algorithms, inappropriate robotic systems and freak accidents. There are always challenges with the adoption of new technologies and someone has gone ahead and collated all the tech horror stories in one place. We’re techno-optimists but it's important to understand the risks if you’re going to avoid them.

Liquid metal leech robots - Researchers have finally developed the shape-shifting tech used by Terminator 2’s villain. The engineers were inspired by leeches and their ability to move flexibly in 3D when responding to external stimuli. They copied this concept and developed a droplet-based soft robot that can conduct electricity, re-shape in response to magnetic fields, and respond to infrared light.

Amazon ignores the FTC’s anti-trust concerns -  Chair of the FTC, Lina Khan expressed concern over Amazon’s acquisition of iRobot. She argued that the juggernaut's willingness to sustain losses, expand across multiple business lines and control data created market dominance and traditional antitrust arguments ignore the potential harm. Meanwhile, Amazon keeps hoovering up companies.

Library of American dynamism - We’ve spoken about the American Dynamism trend in VC at length. Check out this database of every single article on the topic - it's a useful reference and I’m sure could bring value to prospecting salespeople (I know you’re out there). Having lived through many VC buzzword bingo hype cycles it will be interesting to see what the scene looks like after the dust settles.

Artificial general intelligence and Doom -The founder of Doom and software OG John Carmack has raised funding for his stealth mode AGI company. Carmack has a ridiculous resume - he pioneered 3D graphics and computer games in the 80s, he was heavily involved in Oculus before its purchase by Meta and is repeatedly voted one of the most influential people in tech. Watch this space. Side note, its interesting how well suited game developer are to AI e.g Deepmind.

Allergic to restaurant robots - This fairly cynical article argues that hospitality robots are all hype. It states that robots cost more to employ than human servers, due to maintenance and once the novelty has worn off customers want to see a friendly human face. Time will tell.

The Big Idea

Drawing with deep learning robots

We’re currently being spoiled by a glut of research papers creatively applying Deep Reinforcement Learning (DRL) on robotics challenges. Understanding this technology is becoming a necessary requirement for those in the field. Next week -  we will be diving deep into DRL for robotics so think of this week as a little teaser.

From Scratch to Sketch details how a DRL approach trained a cobot to draw. We’ve mentioned a few times that Remix is currently working on a painting robot, so anything arts and crafts get us excited. Their framework allowed the system to autonomously develop its own drawing strategy over time, through a process of trial and error. Its a clever project and we like the lead researchers thinking-

“The primary motivation for our research was to make something cool”