Big Tech &
SoftBank Notches a $6.5 Billion Loss on Plunging WeWork and Uber Valuations (2 minute read)
Masayoshi Son has disclosed the extent of SoftBank Group Corp.'s losses on WeWork and Uber Technologies Inc. SoftBank reported its first quarterly operating loss in 14 years at around $6.5 billion, with $4.6 billion in losses attributed to WeWork. Son's $100 billion Vision Fund had contributed over $14 billion in mostly paper gains over the past two years, but its high exposure to businesses that prioritize growth over profitability is starting to have a negative effect. Despite the hits, SoftBank has still made money for its investors. SoftBank continued to invest more money into WeWork in a deal that included making WeWork's founder Adam Neumann resign from the company's board. Analysts are concerned with Son's investment strategy due to WeWork's poor performance.
Tesla will unveil its Cybertruck pickup on Nov. 21 in LA, Elon Musk says (1 minute read)
Elon Musk has announced that Tesla's Cybertruck will be unveiled on November 21 in Los Angeles, the same location and month as in the opening credits of the movie 'Blade Runner'. The truck includes features such as a power outlet that will allow the use of heavy-duty 240V high power tools in the field all day without the use of a generator. Tesla currently sells three vehicles: the Model S, Model 3, and Model X. It is planning to release the Model Y by the end of 2020. A Tesla Roadster and a Semi truck have been planned for release after the Model Y. Tesla delivered a profit in Q3. It is ahead of schedule for its new factory in Shanghai and its Model Y crossover. Tesla will start manufacturing cars at its European Gigafactory in 2021.
EXCLUSIVE: This Is How the U.S. Military’s Massive Facial Recognition System Works (7 minute read)
The US military has been building a massive database with millions of images of faces, irises, fingerprints, and DNA data over the last 15 years. There are at least 7.4 million identities in the Automated Biometric Information System, including the identities of suspected terrorists and allied soldiers training with the US forces. The Defense Forensics and Biometrics Agency is tasked with overseeing the database. This information was used thousands of times in the first half of 2019 to identify non-US citizens on the battlefield. It is linked to the FBI's biometric database, so the US military can search against the biometric data of US citizens as well. This network will ultimately amount to a global surveillance system. The US military has invested over $345 million in biometric database technology in the last 10 years. Much about the system is still unknown, including its accuracy rate at identifying individuals correctly.
UPS delivers prescription medications to US homes by drone for the first time (1 minute read)
UPS Flight Forward has completed the first prescription medication drone delivery in the US. The company flew the medication from a pharmacy to a home in North Carolina before making a second delivery to a nearby retirement village. UPS partnered with CVS Pharmacy and Matternet for the deliveries. The Matternet M2 drone flew autonomously, but it was monitored by a remote operator. Other companies have completed drone deliveries before, but UPS Flight Forward's service is the first to include prescription medicine. Drone delivery progress has been slow mainly due to regulations that prevent companies from flying drones without FAA authorization.
Programming, Design & Data Science
ipv4.cat is a server that replies with a remote address on all TCP ports. It can be used to determine your public IP address from the command line. ipv4.cat can be used to check for outbound firewall blocks. It returns a raw TCP response, which means it doesn't require an HTTP client.
Building a Feedforward Neural Network from Scratch in Python (18 minute read)
A feedforward neural network, also known as a Multi-layered Network of Neurons, is a network of models where information only travels forward through input nodes and output nodes. Traditional models such as the Perceptron are limited to linear functions. Multilayered-Network of Neurons are able to handle the complex decision boundary between inputs and outputs. This article is a guide on how to build and test a simple neuron network from scratch.
Remember the Uber self-driving car that killed a woman crossing the street? The AI had no clue about jaywalkers (3 minute read)
The first recorded death by a fully autonomous vehicle happened in March 2018, when a self-driving Uber car hit and killed a jaywalking pedestrian. An investigation into the accident found that the code didn't recognize the person as a pedestrian as they did not cross at an obvious designated crossing. The car had detected their presence, but it ignored them, classifying the person as a non-threatening object. At the time of the accident, the vehicle had a human driver at the wheel, but the driver was looking away from the road moments before the accident. The National Safety Transportation Board is set to hold a public meeting on November 19 to determine the probable cause of the collision. Uber will continue to improve the overall safety of their system by training it in scenarios that involve jaywalking, hard braking, and unprotected turns.