Elon Musk unveils Tesla Bot, a humanoid robot that uses vehicle AI

by Moalla Ilani

“It’s intended to be friendly,” the carmaker’s CEO joked.

Jackson Ryan
Aug. 20, 2021 5:00 a.m. PT

Tesla CEO Elon Musk on Thursday unveiled a humanoid robot called the Tesla Bot that runs on the same AI used by Tesla’s fleet of autonomous vehicles. A functioning version of the robot didn’t make an appearance during Musk’s reveal, though a slightly bizarre dance by a performer dressed like a Tesla Bot did.

The unexpected reveal came at the end of Tesla’s AI Day presentation, with Musk providing few details about the slightly creepy, Slenderman-like robot beyond a few PowerPoint slides. The 5-foot-8-inch robot is expected to weigh in at 125 pounds and be built from “lightweight materials,” he said.

Its head will be kitted out with the autopilot cameras used by Tesla’s vehicles to sense the environment and will contain a screen to display information. Internally, it will be operating via Tesla’s Full Self-Driving computer.

“It’s intended to be friendly,” Musk joked, “and navigate through a world built for humans.”

The robot’s appearance came after a 90-minute presentation detailing some of the artificial intelligence upgrades driving Tesla’s electric vehicles, including the Dojo supercomputer, which helps train cars to navigate city streets without human assistance. “It makes sense to put that onto humanoid form,” Musk said.

Three slides detailed the robot’s proposed specifications, and Musk made sure he pointed out that you could both outrun the Tesla Bot and “overpower” it. He has, in the past, railed against the use of robots as weapons and warned of the risks AI might pose — once calling it the “biggest risk we face as a civilization.” I guess if they’re your incredibly slow, easy-to-overpower robots, the dangers are reduced.

“We should be worried about AI,” Musk reiterated during a question and answer session after the presentation. “What we’re trying to do here at Tesla is make useful AI that people love and is … unequivocally good.”

One particular slide said the Tesla Bot would eliminate “dangerous, repetitive, boring tasks,” and Musk provided an example, suggesting the robot could be told to “go to the store and get … the following groceries.” Not that such a task is particularly dangerous, but you might find it repetitive and boring.

Musk, prone to making bold statements about the future, riffed a little on how he envisions Tesla Bot changing the future of work, too. “This, I think, will be quite profound,” he said. “Essentially, in the future, physical work will be a choice. If you want to do it, you can, but you won’t need to do it.”

It’s hard to say how far off such a future might be, but there’s a huge gap between showing off a few PowerPoint slides and delivering an actual, working humanoid robot. It’s probably going to be a long while before you get your bread and milk via the Tesla Bot, but, Musk said, a prototype will likely be ready next year.

https://www.cnet.com/news/elon-musk-unveils-tesla-bot-a-humanoid-robot-utilizing-vehicle-ai/

Sphero’s cute car-shaped robot is driven to teach kids about programming

by Moalla Ilani

It includes puzzles that require logically giving computers instructions
By Mitchell Clark May 19, 2021, 11:34pm EDT

Sphero, a long-time maker of robotic toys, has announced a robotic toy car that’s designed to teach young children about the fundamentals of programming (via Gizmodo). It’s called the Sphero indi, and kids can use different-colored silicone tiles to give the car instructions, letting them create courses and mazes (and hopefully learn how to logically solve problems while doing so).

The indi uses a color-sensor to check which tile it’s driving over, with green tiles telling it to speed up, pink tiles telling it to turn left, purple tiles telling it to stop and celebrate, and so on. To teach kids how to create instructions to get indi from point A to point B, the toy comes with challenge cards that include patterns with missing tiles, so they’ll have to figure out which color tile will help the car reach its goal.

The indi can follow the tiles’ instructions without any sort of connection to a phone or computer, but if your kids are ready for a bit more control, the Sphero Edu Jr app will let them use a block-based language to customize the car’s behavior.

As someone who’s actually taught children programming essentials, block-based coding and having kids design paths for a character to follow all sounds very familiar: MIT’s Scratch programming language gives students a playground to figure out how computers use logic, with many exercises involving solving puzzles and mazes with code. However, I also know that watching a character move on screen isn’t as exciting to kids as watching a real-life toy move through the house. Plus, a sprite bumping into a virtual wall isn’t nearly as funny as watching a toy crash into something or roll off a kitchen table.

The indi is currently available for pre-order, with Sphero estimating that shipping will start in September. An individual student kit, which includes the car, a case, 20 tiles, and 15 challenge cards, costs $125, while a class set that includes materials for eight students and includes a bulk charging case costs $1,200.

Robotic toys for educational purposes are nothing new: the Cozmo robot, formerly made by Anki, and now being iterated on by Digital Dream Labs, used a Scratch-based programming language to put kids in control of a little tractor-like robot, Lego has its Boost and Mindstorms robot kits, and Fisher-Price has a somewhat terrifying Code-a-pillar which has kids add and remove segments from a bug to give it instructions. Sphero itself even has its Mini robot, which has more play-style options, at just $50.

However, like most of the other toys, the Sphero Mini (and its bigger brother the Bolt) requires the use of a tablet, phone, or computer, which the indi doesn’t. For parents looking to get their kids away from the iPad for a bit, the indi could be a way to do that, while still giving them a learning experience.

 

 

Source:

Mitchell Clark May 19, 2021, 11:34pm EDT

Microsoft Flight Simulator is an impressive Xbox Series X workout

by Moalla Ilani

It’s out on Xbox Game Pass tomorrow
By Sam Byford@345triangle Jul 26, 2021, 5:50am EDT

Nothing has pushed my PC further than Microsoft Flight Simulator. That might sound odd for a “game” that’s largely about flying around empty skies by yourself, but Asobo’s latest iteration of the classic franchise is technically groundbreaking and ambitious, with all manner of wizardry going on beyond the scenes to stream in accurate city data, real-time weather effects, and so on. I’ve still had a great time with it, but compared to most AAA games, Flight Simulator asks a lot more of your CPU.

That’s why I was intrigued by the new version for Xbox Series consoles, which comes out on Game Pass tomorrow. In fact, it’s the first Microsoft game for Xbox Series consoles that won’t run natively on any Xbox One model at all, though an xCloud version is also coming to mobile and will eventually hit older Xbox consoles as well. I’ve been playing a preview build provided by Microsoft on my Series X for a few days, and as with my PC, I think it’s the strongest workout for the hardware so far.

Microsoft Flight Simulator is essentially the same proposition on Xbox as it is on PC, offering players the chance to pilot a variety of aircraft around a beautifully rendered version of our planet. The tutorials have been tweaked a little, with a series of shorter missions that should make it easier to get up to speed, and the various commands have been mapped to the Xbox controller in a straightforward, accessible way. You can make the flight model just as complex as the PC version if you want, although right now there aren’t a lot of Xbox-compatible flight stick options.

On the Series X, Flight Simulator runs at 4K resolution and targets 30 frames per second. Overall, I got a more stable experience on the Series X than my own i5 6600K/GTX 1080 rig, which was impressive when I put it together five years ago but somewhat less so now. The frame rate isn’t perfectly smooth — you can drop below 30 when flying low in dense areas like downtown Manhattan, for example, and that’s noticeable. It helped that I played on an LG CX OLED TV, which is capable of variable refresh rates and means you don’t experience tearing or stuttering when the frame rate does fluctuate above or below 30.

Graphical settings are broadly comparable to what you’d get on a good gaming PC, if not quite at the top of the line. The game consistently looks stunning when you’re high in the air, and any seams in the experience are only really apparent when flying close to the ground. That tends to be more to do with how the photogrammetry streaming technology works — again, if you fly quickly into Manhattan or Shinjuku, not every skyscraper is always going to be loaded into memory at once, meaning some buildings might appear a little wobbly at first. I also noticed a few amusing glitches from time to time, like cars driving on the surface of the River Thames in London instead of on Tower Bridge directly above.

As for the Series S, my colleague Tom Warren has spent some time testing that version, and the results are impressive for a tiny $299 box. The game runs at 1080p with reduced graphical effects and draw distances, but as you’ll see from the video, it delivers a solid Flight Simulator experience and will be by far the cheapest way to achieve it.

Microsoft Flight Simulator has improved a lot since its launch last year, with “world updates” that expand the more detailed photogrammetry data further across the globe. That’s all there in the Xbox version, too, including the most recent Nordics update that includes hand-rendered airports and points of interest across Scandinavia, Iceland, and Finland. (It’s also worth noting that the PC version is getting a further update this week that Microsoft promises should dramatically improve performance across the board — stay tuned for how that works in practice.)

If anything, the Xbox version can feel a little too close to the PC version at times, with an occasionally clunky cursor-driven interface. It’s a little conspicuous to have a graphic settings menu where the only option is to turn HDR on and off, for example. But it’s better to leave too much in than to cut too much out: what matters is that the flying experience is as good as it could be given the hardware on hand.

From what I’ve played of Microsoft Flight Simulator on the Xbox Series X, I don’t feel like Asobo has left much on the table. It’s still an incredible technical achievement, and one well worth checking out when it hits Game Pass tomorrow.

The robotic farm of the future isn’t what you’d expect

by Moalla Ilani

Iron Ox’s robots grow leafy greens in hydroponic vats, but humans are definitely needed
By James Vincent Oct 9, 2018, 10:05am EDT

When we think about automation, we often imagine robots just doing the work of humans. Our mental image is of an android in overalls, clocking in with a lunchbox full of oil and bolts, and grabbing a hammer. But that’s not what happens. The reality is much messier, and the process of automation is one of compromise and incremental progress.

Agritech startup Iron Ox is the perfect example of this. After launching in 2015 with the aim of automating the hard work of growing produce, the company unveiled its first “autonomous” production farm last week. In 8,000 square feet of indoor space (roughly 0.2 acres), its engineers use proprietary robot systems to grow roughly 26,000 heads of lettuce, leafy greens, and herbs each year in hydroponic vats. The company says it should start selling its crop in “the next couple of months,” and it will be targeting restaurants first.

But, as co-founder and CEO Brandon Alexander explains to The Verge, despite the “autonomous” moniker, humans still play crucial roles in this farm. They are the ones who plant each seedling and package the finished product. Robots just aren’t ready to do it all.

Getting to even this level of automation took years of work to navigate the constraints of modern robotics, says Alexander, who previously worked at Google’s research lab X and at influential robotics incubator Willow Garage. “There’s a huge difference between your equipment working one time for a video and it working every day,” he says. “Most people outside of robotics underestimate just how big that gap is.”

Hydroponics has a lot of advantages over traditional farming. It uses less water, it’s completely sterile, and it’s much more space-efficient. (Iron Ox claims it grows 30 times more produce per acre than a regular farm.) But it also requires more labor. In the sort of hydroponics system that Iron Ox uses, each plant has to be transferred one by one through a series of different growing vats based on its rate of growth. It’s this simple task — moving plants — that has been the focus of Iron Ox’s automation efforts.

“The bulk of the cost [in hydroponic farming] is all those touchpoints,” says Alexander. “Labor is over 50 percent of the cost to grow right now. If you’re indoors, there are no tractors, no nothing, so it’s a very manual product.”

To overcome this, Iron Ox developed two robotic systems. The first is a porter: a 1,000-pound wheeled bot named Angus that moves pallets of seedlings around the warehouse in their hydroponic growing vats. The second is a robotic arm (currently unnamed) that picks up individual plants and moves them from vat to vat. All of this is overseen by a computer program (nicknamed “the Brain”) that monitors the growing conditions in each pallet and adjusts their balance of gases and nutrients for optimal growing conditions.

Alexander says it was a huge amount of work just to get to this level, reeling off a list of problems the team faced. There was the lighting, he says, as Iron Ox had initially planned to use all LEDs for growing. “But that results in electricity bills that are simply too expensive. So, going forward, we’re betting on more traditional greenhouses augmented by LEDs.”

There was also a myriad of robotic challenges in developing things like leaf detection for the robot arm so it doesn’t drag plants through one another when moving them, and designing Angus to accelerate and decelerate at gentle rates so it didn’t slop growing medium out of the vats. “The best motivator for engineers to solve that problem was having to clean up the water afterward,” says Alexander.

Overcoming these technical problems was tricky, but it’s nothing compared to the challenge that the company now faces: cost. If Iron Ox truly wants to change farming, it needs to compete on price with modern outdoor farms that have been optimized over the past century to produce incredible amounts of produce as cheaply as possible.

Jonathan Gill, a robotics researcher at Harper Adams University and co-founder of the Hands Free Hectare, an automated farming project in the UK, says this will be difficult. Only “minimal amounts” of the global agricultural industry have been automated to date, says Gill, simply because it’s not “economically viable” to do so. The robots aren’t good enough, human labor is cheaper, and those equations aren’t going to change just for Iron Ox, he says. “It’s just not necessarily the most efficient method of growing food for the world.”

But, says Gill, Iron Ox has other advantages. “They’re always going to be able to market it as a premium product,” says Gill. “[Their produce] is completely organic, has little to no human input, and they can claim it’s fresher and has fewer food miles than competitors.”

They can also grow closer to customers and adapt their crops to last-minute changes in demand. “Here in the UK, we had a very hot summer, and as soon as it passed 30 degrees [Celsius], the lettuce stopped growing,” says Gill. Supermarkets went through a salad shortage, but in that sort of situation, Iron Ox will be able to “guarantee a market push-through for large customers.”

Alexander mentions the same benefits and says Iron Ox farms will be located near urban centers and will easily adapt to the whims of the market. But he also claims they’ll compete with big farms on price “from day one.” This first farm will sell greens to small customers first, but Alexander claims they’ll be selling to big chains by 2019.

“Our ultimate goal is not to sell produce to bougie grocery stores,” says Alexander. “[Our rivals] are huge outdoor fields, and if we can’t compete with them, we’re always going to be a niche player.”