USB Over Ethernet enables sharing of multiple USB devices over Ethernet, so that users can connect to devices such as webcams on remote machines anywhere in the world as if the devices were physically plugged into their own computers. The flaws are in the USB Over Ethernet function of the Eltima SDK, not in the cloud services themselves, but because of code-sharing between the server side and the end user apps, they affect both clients – such as laptops and desktops running Amazon WorkSpaces software – and cloud-based machine instances that rely on services such as Amazon Nimble Studio AMI, that run in the Amazon cloud. The flaws allow attackers to escalate privileges so that they can launch a slew of malicious actions, including to kick the knees off the very security products that users depend on for protection. Specifically, the vulnerabilities can be used to “disable security products, overwrite system components, corrupt the operating system or perform malicious operations unimpeded,” SentinelOne senior security researcher Kasif Dekel said in a report published on Tuesday.
When we first looked at the idea of Rust in the Linux kernel, it was noted that the objective was not to rewrite the kernel’s 25 million lines of code in Rust, but rather to augment new developments with the more memory-safe language than the standard C normally used in Linux development. Part of the issue with using Rust is that Rust is compiled based on LLVM, as opposed to GCC, and subsequently supports fewer architectures. This is a problem we saw play out when the Python cryptography library replaced some old C code with Rust, leading to a situation where certain architectures would not be supported. Hence, using Rust for drivers would limit the impact of this particular limitation. Ojeda further noted that the Rust for Linux project has been invited to a number of conferences and events this past year, and even garnered some support from Red Hat, which joins Arm, Google, and Microsoft in supporting the effort. According to Ojeda, Red Hat says that “there is interest in using Rust for kernel work that Red Hat is considering.”
DeepMind, which regularly feeds its work into Google products, has probed the capabilities of this LLMs by building a language model with 280 billion parameters named Gopher. Parameters are a quick measure of a language’s models size and complexity, meaning that Gopher is larger than OpenAI’s GPT-3 (175 billion parameters) but not as big as some more experimental systems, like Microsoft and Nvidia’s Megatron model (530 billion parameters). It’s generally true in the AI world that bigger is better, with larger models usually offering higher performance. DeepMind’s research confirms this trend and suggests that scaling up LLMs does offer improved performance on the most common benchmarks testing things like sentiment analysis and summarization. However, researchers also cautioned that some issues inherent to language models will need more than just data and compute to fix. “I think right now it really looks like the model can fail in variety of ways,” said Rae.
The Great Resignation is real, and it has affected the logistics industry more than anyone realizes. People don’t want low-paying and difficult jobs when there’s a global marketplace where they can find better work. Automation will be seen as a way to address this, and in 2022, we will see a lot of tech VC investment in automation and robotics. Some say SpaceX and Virgin can deliver cargo via orbit, but I think that’s ridiculous. What we need, (and what I think will be funded in 2022, are more electric and autonomous vehicles like eVTOL, a company that is innovating the “air mobility” market. According to eVTOL’s website, the U.S. Department of Defense has awarded $6 million to the City of Springfield, Ohio, for a National Advanced Air Mobility Center of Excellence. ... In 2022 transformations, grocery will cease to be an in-store retail experience only, and the sector will be as virtual and digitally-driven as the best of them. Things get interesting when we combine locker pickup, virtual grocery, and automated last-mile delivery using autonomous vehicles that can deliver within a mile of the warehouse or store.
In a broad sense, a penetration test works in exactly the same way that a real attempt to breach an organization's systems would. The pen testers begin by examining and fingerprinting the hosts, ports, and network services associated with the target organization. They will then research potential vulnerabilities in this attack surface, and that research might suggest further, more detailed probes into the target system. Eventually, they'll attempt to breach their target's perimeter and get access to protected data or gain control of their systems. The details, of course, can vary a lot; there are different types of penetration tests, and we'll discuss the variations in the next section. But it's important to note first that the exact type of test conducted and the scope of the simulated attack needs to be agreed upon in advance between the testers and the target organization. A penetration test that successfully breaches an organization's important systems or data can cause a great deal of resentment or embarrassment among that organization's IT or security leadership
According to Bharadwaj, the concrete and steel environment effectively acted as a “Faraday cage,” which meant that the EV chargers wouldn’t talk to people’s mobile phones when they tried to initiate charging. You could find yourself stranded, unable to charge your car. “So we had to innovate.” ... As with any EV charging, a payment app connects your car to the EV charger. With Xeal, the use of NFC means the only time you need the Internet is to download the app in the first instance to create a profile that includes their personal and vehicle information and payment details. You then receive a cryptographic token on your mobile phone that authenticates your identity and enables you to access all of Xeal’s public charging stations. The token is time-bound, which means it dissolves after use. To charge your car, you hold your phone up to the charger. Your mobile reads the cryptographic token, automatically bringing up an NFC scanner. It opens the app, authenticates your charging session, starts scanning, and within milliseconds, the charging session starts.
The scarcity of skilled AI developers or engineers stands as a major barrier to adopting AI technology in many companies. No-code and low-code technologies come to the rescue. These solutions aim to offer simple interfaces, in theory, to develop highly complex AI systems. Today, web design and no-code user interface (UI) tools let users create web pages simply by dragging and dropping graphical elements together. Similarly, no-code AI technology allows developers to create intelligent AI systems by simply merging different ready-made modules and feeding them industrial domain-specific data. Furthermore, NLP, low-code, and no-code technologies will soon enable us to instruct complex machines with our voice or written instructions. These advancements will result in the “democratization” of AI, ML, and data technologies. ... In 2022, with the aid of AI and ML technologies, more businesses will automate multiple yet repetitive processes that involve large volumes of information and data. In the coming years, an increased rate of automation can be seen in various industries using robotic process automation (RPA) and intelligent business process management software (iBPMS).
Large language models like OpenAI’s GPT-3 show an aptitude for generating humanlike text and code, automatically writing emails and articles, composing poetry, and fixing bugs in software. But the dominant approach to developing these models involves leveraging massive computational resources, which has consequences. Beyond the fact that training and deploying large language models can incur high technical costs, the requirements put the models beyond the reach of many organizations and institutions. Scaling also doesn’t resolve the major problem of model bias and toxicity, which often creeps in from the data used to train the models. In a panel during the Conference on Neural Information Processing Systems (NeurIPS) 2021, experts from the field discussed how the research community should adapt as progress in language models continues to be driven by scaled-up algorithms. The panelists explored how to ensure that smaller institutions and can meaningfully research and audit large-scale systems, as well as ways that they can help to ensure that the systems behave as intended.
While firms have narrowed their scope to address more targeted pain points, the increased digitalisation of assets is helping to drive interest in the adoption of DLT in new ways. Previous talk of mass disruption of the financial system has given way to more realistic, but still transformative, discussions around how DLT could open doors to a new era of business workflows, enabling transactional exchanges of assets and payments to be recorded, linked, and traced throughout their entire lifecycle. DLT’s true potential rests with its ability to eliminate traditional “data silos”, so that parties no longer need to build separate recording systems, each holding a copy of their version of “the truth”. This inefficiency leads to time delays, increased costs and data quality issues. In addition, the technology can enhance security and resilience, and would give regulators real-time access to ledger transactions to monitor and mitigate risk more effectively. In recent years, we have been pursuing a number of DLT-based opportunities, helping us understand where we believe the technology can deliver maximum value while retaining the highest levels of risk management.
Simple is often better: You can do (almost) anything with technology, but it doesn't mean you should. Especially in the security space, many customers overengineer solutions. I like this video from Google’s Stripe conference to underscore this point. People, process, technology: Design for people to enhance process, not tech first. There are no "perfect" solutions. We need to balance various risk factors and decisions will be different for each business. Too many customers design an approach that their users later avoid. Focus on 'why' first and 'how' later: Be the annoying 7-yr old kid with a million questions. We can't arrive at the right answer if we don't know the right questions to ask. Lots of customers make assumptions on how things need to work instead of defining the business problem. There are always multiple paths that can be taken. Long tail of past best practices: Recognize that best practices are changing at light speed.
Quote for the day:
"Eventually relationships determine the size and the length of leadership." -- John C. Maxwell