The ability to pass down knowledge from one to another through a repetitive teaching reward system has proven to be one of the pivotal aspects of human life. An example of this would be teaching a dog to sit or roll over, then to reward it with a treat after completion of the desired task.
This basic principle practice increases the likelihood of the dog performing the same task more efficiently once learned that their is a reward after.
The same principle practice can now be applied to AI robots, being named one of the 10 breakthrough technologies of 2017 by MIT Technology Review, this revolutionary kind of AI learning will mean that AI will be able to adapt and take in new information without human interaction, simply by experimenting.
SpaceX has added a newly updated funding round to this year's summer collection. Adding $100 million on to it's already $351 million funding round bringing the total to around $450 million, putting Space X's total post-money valuation to $21.5 billion.
James Gleeson, the spokesperson for SpaceX has replied to the amended filing by saying: "This filing simply discloses that SpaceX received an additional $100 million of investment as part of the last funding round which was disclosed earlier this summer."
SpaceX has massive ambitions, outlining plans for a trip to Mars in 2022 and launching satellites for the U.S military using its Falcon 9 rocket. It's clear why investors see SpaceX as a trustworthy investment, the track record for Elon provides its benefits of reassurance.
Quantum encryption is the wish of most in the security industry as you can't even look at the data without making changes to it, but using quantum encryption - until now - was always slower than traditional methods.
Not for long, as researchers have made a new quantum encryption key distribution system that is between 5-10x faster than previous methods, and can keep up with conventional encryption when run side-by-side. How did they do it? By getting more data into each photon.
Engadget reports "Normally, you can only encode one bit per photon by using a weak laser. The team discovered that it could encode two bits by tweaking the release time of photons and using high-speed photon detectors to track these changes. Effectively, they're giving photons properties they couldn't have before". Crazy stuff, isn't it?
Elon Musk, or the real life Tony Stark, has taken the first step into solving the standstill traffic problem that LA is known for. Musk has begun the process with his firm 'The Boring Company' by filing a permit to start digging within city limits.
Keeping the momentum going forward Musk has already acquired a verbal approval and has also posted a photo from the first stage of the project near his office in Hawthorne, in southwestern Los Angeles County near LAX. Musk hopes the route of tunnel will stretch from the entirety of the "whole 405" from LAX to Interstate 101 in "a year or so."
In April, The Boring Company released a video showing a car being lowered into the tunnel from a platform in the city streets. Users of the tunnel will be able to take "express" trips through the main tunnel with "no intermediate stopping", side tunnels will also be connected to the main tunnel for entry and exit.
The website states: "Therefore, unlike trains, the skate's average speed is very close to its maximum speed. And with no intermediate stopping, there is no practical upper limit to the number of stations along the alignment, and the stations can be built as small as a single parking space. Each skate can carry either a single automobile or 8-16 passengers (mass transit) based on configuration."
Earlier this year the state-funded Russian Foundation for Advanced Research Projects unveiled the Final Experimental Demonstration Object Research - or better known as FEDOR. The robot demonstrated the ability to fire pistols correctly at targets, crawl and even drive a car.
In a tweet from Russian Deputy Prime Minister and Deputy Head of the Military-Industrial Commission Dmitry Rogozin said "Robot FEDOR showed the ability to shoot from both hands. Fine motor skills and decision-making algorithms are still being improved."
FEDOR is now having its decision making algorithms upgraded as the robot is destined to orbit the Federatsiya in 2021. Cosmonauts plan to use this very same spacecraft to hopefully carry astronauts to the moon one day, and also aspire to travel to Mars it the FEDOR as well.
Great, this is just what we needed: robotic dogs. Boston Dynamics has been hard at work making some of the scariest robots outside of the movies, with the company showing off the latest version of their SpotMini robotic dog.
The new SpotMini dog looks much cleaner and more '2017-y' as the firm has used a proper body on the robotic dog, with movements that are scarily real. It seems very quiet, so you won't hear it sneaking up on you, which I don't know whether that's a good, or a bad thing. I'm still undecided on that, as robotic dogs are freaky as hell.
The cable that is providing millions of people with internet access, and even the cable you have running to your router - heck, the same fiber optic cable that is going into my router here as I type this, could soon be an early earthquake detector.
Researchers have developed a new technology that detects seismic activity through the "jiggling" in fiber optic cables, with 'laser interrogators' looking for disturbances in the fiber, and then sending that information about the magnitude and direction of the tremors. Using fiber-based detections isn't something that's completely new, but it has previously been used for acoustic sensing that needed them wrapped in cement,or sticking them to a surface so that you ensured contact was made to the ground.
But this new system is capable of detecting different types of seismic waves, so it would know if the impending earthquake was something worth alerting someone, or a system somewhere about it. The new system is also capable of detecting very small, or localized quakes that would otherwise not be noticed.
TSMC is ramping into 3nm as soon as possible, with the Taiwanese semiconductor giant spending $20 billion on a next-gen 3nm manufacturing plant, so that it can keep customers like Apple happy in the future.
Samsung and Intel have been in the fab game for a while now, but TSMC has been spending around $10 billion per year in order to keep up to, or exceed what their two main competitors are doing. TSMC has announced its building a next-gen 3nm technology plant in Southern Taiwan, but to put it into perspective... 1nm is one billionth of a meter, and measures the width of each transistor in the chip. It's an incredibly intricate process, with TSMC leading the world in many ways.
Morris Chang, founder of TSMC and 86-year-old ultimate boss of the company said: "By the time we're through, by the time we've built all the necessary capacity, I think we would have spent upwards of $15 billion. That's a conservative estimate. Maybe it's safer to say upwards of $20 billion". Don't worry, TSMC won't be going out of business any time soon, as his company is worth a whopping $190 billion.
TSMC's new 3nm plant will be ready by 2022, so expect next-gen Apple smartphones and tablets to rock TSMC's upcoming 3nm node sometime after 2022. 3nm will usher in a 3x reduction in semiconductor size, giving us faster processors that use less power. We should expect quite a leap in smartphone battery life with 3nm, which is the part I'm looking forward to most.
OMRON has been making a table tennis robot for a while now, with the AI-powered future table tennis champion dubbed FORPHEUS in its fourth-generation state now able to use its new companion robot arm that throws the ball up in the air so that it can serve.
This means that FORPHEUS is a great trainer and/or opponent for would-be table tennis players, with the fourth-gen robot featuring a display on both sides of the "net" that display players' status.
There's also a three-camera system that faces the human player that is built into the main body, and a bat that is driven by three arms and a 5-axis motor.
All of this is run by a huge motion controller computer that is behind the stage.
TSMC is running directly into the arms of 3nm fabrication, with the Taiwanese semiconductor company announcing the locaiton of their first 3nm fab, which will be build in Tainan Science Park, in southern Taiwan.
The decision to build it in Taiwan versus the rumors of a US production plant is so that TSMC can use their assets and supply chain for the production and support of many customers like AMD, NVIDIA, who will want 3nm designs made. The Taiwanese government also provided a lending hand, with land, water, electricity and environmental protection offered to TSMC for their latest manufacturing plant. Some of the new 3nm manufacturing machines will be supplied by ASML, a Netherlands-based company.
As for when we can expect 3nm production to kick off... 2022 at the earliest. TSMC has teased that the 7nm yields are currently ahead of schedule, so we should expect 7nm-based products released in 2018. The 5nm roadmap is still on track, expected to launch in Q1 2019.