Daily Tech Digest - January 18, 2021

Go back to the office? Some employees would rather quit

It's been difficult to get a good read on what the prevailing attitude is towards working from home since traditional workplaces shut in the early months of 2020. While the consensus largely appears to be that employees relish the flexibility and (subjective) comfort that working from home provides, this is at odds with the mental health spectre that has loomed over the COVID-19 crisis, with countless reports and statistics highlighting the toll that working in insolation has on wellbeing. This is captured in LiveCareer's survey. Despite 81% of employees saying they enjoyed working remotely, and 61% expressing a desire to continue working in a remote capacity after the pandemic is over, only 45% of those polled said that telecommuting had not taken a toll on their mental wellbeing. Clearly, the ideal situation is about balance: employees want the option to work remotely, while also having a shared workspace that they can use when needed. One major factor employers need to address to make the return to office life more appealing is safety. With many companies still trying to figure out how they can reconfigure or otherwise re-think their real estate investments to suit a new hybrid workforce, ensuring workplaces are safe should top the agenda.

AVIF Image Format: The Next-Gen Compression Codec

AVIF, or AV Image Format, is an open-source and royalty-free image format based on the AV1 codec, and ,similar to AV1, AVIF provides a very high compression rate. The fact that it's royalty-free makes it stand out from the competition. Leveraging the power of AV1 has proven beneficial for AVIF, in both processing time and its ability to handle hardware issues. Before we further discuss the advantages of AVIF, be advised that AVIF saves pictures in AVIF image format, which is relatively new and still not widely adopted. On top of this, it's using a reasonably new algorithm. So there may be a possibility that it’s not best for all use cases right now. ... The idea behind designing AV1 was to transmit video over the internet. With a better compression rate for video, AVI reduced the number of overall bits. This allows the AV1 codec to provide multiple coding techniques that gives developers some freedom when writing their code. If you wonder why we brought this concept of video compression technique into an image compression post, its because videos and image codecs share similarities in the nature of their data. The AV1 codec has proved very advantageous for the internet by saving bandwidth, which MPEG could not do, although JPEG XR was still in the race but not as effective as AV1.

Love in the time of algorithms: would you let your artificial intelligence choose your partner?

Another problematic consequence may be rising numbers of socially reclusive people who substitute technology for real human interaction. In Japan, this phenomenon (called “hikikomori”) is quite prevalent. At the same time, Japan has also experienced a severe decline in birth rates for decades. The National Institute of Population and Social Security Research predicts the population will fall from 127 million to about 88 million by 2065. Concerned by the declining birth rate, the Japanese government last month announced it would pour two billion yen (about A$25,000,000) into an AI-based matchmaking system. The debate on digital and robotic “love” is highly polarised, much like most major debates in the history of technology. Usually, consensus is reached somewhere in the middle. But in this debate, it seems the technology is advancing faster than we are approaching a consensus. Generally, the most constructive relationship a person can have with technology is one in which the person is in control, and the technology helps enhance their experiences. For technology to be in control is dehumanising.

How Teams Can Overcome the Security Challenges of Agile Web App Development

Managing company secrets in an agile environment means CISOs need to rethink the scalability of their current security solutions. With rapidly changing codebases, it’s essential that enterprises use security tools that support agile development and also extend to other platforms that devops teams might use. Akeyless is a versatile security tool that fragments encryption keys and provides a high degree of data security. It supports agile release environments and can be scaled to different platforms as needed. One of the reasons I like implementing this solution when consulting for app companies is how easily I can integrate it with all the major development platforms through plugins, ensuring that in-house departments and subcontractors alike can securely manage access to sandbox servers and databases, without interrupting their workflows. Beyond governance concerns, in my experience, compliance and audit teams generally stand to gain a great deal by learning about how automation can help them achieve their goals, along with how their protocols can improve with automation. On the other hand, complete automation might not be possible in every area. 

Multiple backdoors and vulnerabilities discovered in FiberHome routers

FTTH ONT stands for Fiber-to-the-Home Optical Network Terminal. These are special devices fitted at the end of optical fiber cables. Their role is to convert optical signals sent via fiber optics cables into classic Ethernet or wireless (WiFi) connections. FTTH ONT routers are usually installed in apartment buildings or inside the homes or businesses that opt for gigabit-type subscriptions. In a report published last week, security researcher Pierre Kim said he identified a large collection of security issues with FiberHome HG6245D and FiberHome RP2602, two FTTH ONT router models developed by Chinese company FiberHome Networks. The report describes both positive and negative issues with the two router models and their firmware. ... Furthermore, the Telnet management feature, which is often abused by botnets, is also disabled by default. However, Kim says that FiberHome engineers have apparently failed to activate these same protections for the routers' IPv6 interface. Kim notes that the device firewall is only active on the IPv4 interface and not on IPv6, allowing threat actors direct access to all of the router's internal services, as long as they know the IPv6 address to access the device.

How do I select a fraud detection solution for my business?

From strictly rules based to fully black box. The former gives you complete control but can be cumbersome and relies on a knowledgeable in house fraud team. The other end is perfect for extreme transaction volumes but offers little explanability. Fortunately, there is a middle ground with whitebox, supervised machine learning- you get the best of both worlds- granular rule based with machine learning making connections between disparate and complex data points. Fraud detection technologies should fit your business, not the other way around. Fraudsters evolve and find ingenious workarounds to most point solutions. Modern fraud detection is a “net” approach where the latest cutting edge tools are used in combination to make it very, very hard for a fraudster to fool them all. Results are very hard to predict. Try and select fraud technologies that allow you to test and show proof of value with no commitment, free trial periods. Modern, effective fraud tech should follow the best SAAS products where you see actual pricing, monthly contracts and free trials. Product value and risk should rest solely on the fraud detection partner.

The AI Incident Database wants to improve the safety of machine learning

“The goal of the AIID is to prevent intelligent systems from causing harm, or at least reduce their likelihood and severity,” McGregor says. McGregor points out that the behavior of traditional software is usually well understood, but modern machine learning systems cannot be completely described or exhaustively tested. Machine learning derives its behavior from its training data, and therefore, its behavior has the capacity to change in unintended ways as the underlying data changes over time. “These factors, combined with deep learning systems capability to enter into the unstructured world we inhabit means malfunctions are more likely, more complicated, and more dangerous,” McGregor says. Today, we have deep learning systems that can recognize objects and people in images, process audio data, and extract information from millions of text documents, in ways that were impossible with traditional, rule-based software, which expect data to be neatly structured in tabular format. This has enabled applying AI to the physical world, such as self-driving cars, security cameras, hospitals, and voice-enabled assistants. And all these new areas create new vectors for failure.

Chatbot Gone Awry Starts Conversations About AI Ethics in South Korea

Luda came under the national spotlight when it was reported that users were training Luda to spew hate speech against women, sexual minorities, foreigners, and people with disabilities. Screengrabs show Luda saying, “they give me the creeps, and it’s repulsive” or “they look disgusting,” when asked about “lesbians” and “black people,” respectively. Further, it was discovered that groups of users in certain online communities were training Luda to respond to sexual commands, which provoked intense discussions about sexual harassment in a society that already grapples with gender issues. Accusations of personal data mishandling by ScatterLab emerged as Luda continued to draw nationwide attention. Users of Science of Love have complained that they were not aware that their private conversations would be used in this manner, and it was also shown that Luda was responding with random names, addresses, and bank account numbers from the dataset. ScatterLab had even uploaded a training model of Luda on GitHub, which included data that exposed personal information. Users of Science of Love are preparing for a class-action lawsuit against ScatterLab, and the Personal Information Protection Commission, a government watchdog, opened an investigation on ScatterLab to determine whether it violated the Personal Information Protection Act.

Digital transformation: it has never been more relevant for businesses

As it’s not necessarily a tangible metric, it can be hard to measure the return on investment (ROI) of the transformation journey and its success. However, numerous considerations can help determine what the ROI is. Firstly, setting out objectives for transformation – it could be to improve the customer experience, the company’s infrastructure or staff productivity, for example. Secondly, outlining the costs of implementing the transformation strategy is essential – as is knowing what the outcomes of that financial outlay are. This will provide a reference point and clear performance indicators when measuring ROI. Of course, setting realistic goals is important in the first place; stage one of the journey, discovering and assessing, should provide guidance on setting achievable targets. And when implementing new systems, there are different metrics that can be detailed in order to measure their success. For instance, if trying to improve end user experience, tackling common pain points experienced by external parties such as slow load times and application response will help reach the overall goal. If IT systems offer a rapid response, end users won’t feel frustrated by the operating system.

Do you really want a CEO to be a role model?

It’s likely that the effectiveness of role models is rooted in mirror neurons, specialized cells that are located in several areas of the human brain. They were first identified about 30 years ago when neuroscientists who had implanted electrodes in monkeys to study how their brains generated hand movements suddenly realized that the same neurons were firing when the monkeys ate and when the monkeys watched the scientists eat. Since then, some researchers have come to see mirror neurons as the biological mechanism through which humans unconsciously copy the behaviors of others. That conclusion would lend scientific credence to the advice that Sutton got from his dad: Being a jerk can be contagious. The work of sociologist Robert K. Merton offers a clue to avoiding an infection of negative behaviors. Merton made a distinction between role models (a term he coined in the 1950s) and reference individuals. He said that when a person emulates a reference individual, he or she copies that person’s good and bad behavioral traits and values without discrimination. But when a person emulates a role model, the focus is on a more limited segment of behaviors and values. This suggests that you can act like Elon Musk, the entrepreneurial innovator, without becoming Musk, the blurting tweeter.

Quote for the day:

"If one oversteps the bounds of moderation, the greatest pleasures cease to please." -- Epictetus

No comments:

Post a Comment