Daily Tech Digest - August 22, 2021

Move Fast Without Breaking Things in ML

The first step in the response to the problem has happened even before you got invited to the call with your CTO. The problem has been discovered and the relevant people have been alerted. This is likely the result of a metric monitoring system that is responsible for ensuring important business metrics don’t go off track. Next using your ML observability tooling, which we will talk a bit more about in a second, you are able to determine that the problem is happening in your search model since the proportion of users who are engaging with your top n-links returned has dropped significantly. After learning this you rely on your model management system to either roll back to your previous search ranking model or deploy a naive model that can hold you over in the interim. This mitigation is what stops your company from losing (as much) money every minute since every second counts for users being served incorrect products. Now that things are somewhat working again, you need to look back to your model observability tools to understand what happened with your model.


Ransomware is the top cybersecurity threat we face, warns cyber chief

Not only are cyber-criminal ransomware groups encrypting networks and demanding a significant payment in exchange for the decryption key, now it's common for them to also steal sensitive information and threaten to release it unless a ransom is paid – often leading victims to feel as if they have no choice but to give in to the extortion demands. "As the business model has become more and more successful, with these groups securing significant ransom payments from large profitable businesses who cannot afford to lose their data to encryption or to suffer the down time while their services are offline, the market for ransomware has become increasingly professional," Cameron will say. Ransomware is successful because it works; in many cases, because organisations still don't have the appropriate cyber defences in place to prevent cyber criminals infiltrating their network in the first place in what the NCSC CEO describes as "the cumulative effect of a failure to manage cyber risk and the failure to take the threat of cyber criminality seriously".


Become software engineers, not software integrators.

Ever since its inception, the IT industry has been evolving every day, by giving better and more awesome technology experiences to end-users. On the other hand, the industry has also continually focused on reducing the development time and cycle for software engineering teams. A significant portion of IT engineers & organizations are motivated to ease the development process. This in turn has become a race to give the best technologies (frameworks, tools, etc.) to engineering teams. In this race, their focus has gradually shifted from “ease of development” to almost “no development at all”, i.e. making tools, which allow the engineers to just integrate stuff to provide the final product. Essentially, plug and play. Of course, the big advantages because of this are that: Now the companies which are building software for businesses can focus more on business ideas; and With a reduced development cycle, companies can build many more software products. However, the concern starts when engineers, who get used to the plug & play tools, start losing core engineering skills like optimizing, maturing, and architecting the code.


How External IT Providers Can Adopt DevOps Practices

The key is to overcome waterfall thinking. A modern supplier will work in small batches and will use an experimental approach to product development. The supplier’s product development team will create hypotheses and valid them with small product increments, ideally in production. According to my experience, many IT suppliers use agile software development and Continuous Integration these days. But they stop their iterative approach at the boundary to production. One problem of having separated silos for development and operations is that in most cases these two silos have different goals (dev = throughput, ops = stability), Diener mentioned. In contrast, a DevOps team has a common business goal. ... In order to adopt DevOps practices, the supplier has to find out what his client’s goal is. It has to become the supplier’s goal as well. We at cosee use product vision workshops to shape and document the client’s goal (impact) and its user’s needs (outcome). That’s a prerequisite for an iterative and experimental product development approach.


Blockchain in Space: What’s Going on 4 Years After the First Bitcoin Transaction in Orbit?

The growth in both scale and affordability of space exploration is creating a whole new sector — the Space Economy, as the United Nations Office for Outer Space Affairs already calls it. An inevitable question then arises: what money will the players in this space economy use? ... Despite all the advances, space exploration often remains a costly business, both in money and science capital. Because of that high cost nature, any large project in space requires the cooperation of numerous private companies, each providing resources and talent. And the most ambitious programs are collaborations between governments — not all of which necessarily put a lot of trust in each other. This is where one of blockchain’s key advantages comes in: it enables the exchange of value and data between independent parties in a way that doesn’t involve trust. With smart contracts, peer-to-peer transaction settlement, and the transparency and accountability enabled by public blockchain records


Upcoming Trends in DevOps and SRE in 2021

Service meshes are quickly becoming an essential part of the cloud-native stack. A large cloud application may require hundreds of microservices and serve a million users concurrently. A service mesh is a low-latency infrastructure layer that allows high traffic communication between different components of a cloud application(databases, frontends, etc.) This is done via application programming interfaces (APIs). Most distributed applications today have a load balancer that directs traffic; however, most load balancers are not equipped to deal with a large number of dynamic services whose locations/counts vary over time. To ensure that large volumes of data are sent to the correct endpoint, we need tools that are more intelligent than traditional load balancers. This is where Service Meshes come into the picture. In typical microservice applications, the load balancer or firewall is programmed with static rules. However, as the number of microservices increases and the architecture changes dynamically, these rules are no longer enough. 


How GPT-3 and Artificial Intelligence Will Destroy the Internet

As a natural language processor and generator, GPT-3 is a language learning engine that crawls existing content and code to learn patters, recognizes syntax and can produce unique outputs based on prompts, questions and other inputs. But GPT-3 is more than just for use by content marketers as witness by the recent OpenAI partnership with Github for creating code using a tool dubbed “Copilot.” The ability to use autoregressive language modeling doesn’t just apply to human language, but also various types of code. The outputs are currently limited, but its future potential use could be vast and impacting. How GPT-3 is Currently Kept at Bay With current beta access to the OpenAI API, we developed our own tool on top of the API. The current application and submission process with OpenAI is stringent. Once an application has been developed before it can be released to the public for use in any commercial application, OpenAI requires a detailed submission and use case for approval by the OpenAI team. 


NFTs, explained

“Non-fungible” more or less means that it’s unique and can’t be replaced with something else. For example, a bitcoin is fungible — trade one for another bitcoin, and you’ll have exactly the same thing. A one-of-a-kind trading card, however, is non-fungible. If you traded it for a different card, you’d have something completely different. You gave up a Squirtle, and got a 1909 T206 Honus Wagner, which StadiumTalk calls “the Mona Lisa of baseball cards.” (I’ll take their word for it.) At a very high level, most NFTs are part of the Ethereum blockchain. Ethereum is a cryptocurrency, like bitcoin or dogecoin, but its blockchain also supports these NFTs, which store extra information that makes them work differently from, say, an ETH coin. It is worth noting that other blockchains can implement their own versions of NFTs. (Some already have.) NFTs can really be anything digital (such as drawings, music, your brain downloaded and turned into an AI), but a lot of the current excitement is around using the tech to sell digital art.


Demystifying AI: The prejudices of Artificial Intelligence (and human beings)

In a way, the results of these algorithms hold a mirror to human society. They reflect and perhaps even amplify the issues already present. We know that these algorithms need data to learn. Their predictions are only as good as the data they are trained on and the goal they are set to achieve. The data needed to train these algorithms is huge (think millions and above). Suppose we are trying to develop an algorithm to identify cats and dogs from pictures. Not only do we need thousands of pictures of cats and dogs, but they should be labeled (say the cat is class 0 and dog is class 1) so that the algorithm can understand. We can download these images off the internet (the ethics of which is questionable), but still, they need to be labeled manually. Now, consider the complexity and effort required to correctly label a million images in one thousand classes. Often this labeling task is done by “cheap labor” who may or may not have the motivation to do it correctly, or they simply make mistakes. Another problem in the data set is that of class imbalance. 


Three Mistakes That Will Ruin Your Multi-Cloud Project (and How to Avoid Them)

A multi-cloud strategy only augments the likelihood of experiencing one of these errors. The complexity of multiple clouds provides an extended attack surface for threat actors. An increased number of services means a higher chance of experiencing a misconfiguration or data leak. Centralized visibility and management are necessary to combat risk and ensure protection and compliance across multi-cloud environments. Proper governance requires a full view of the cloud, complete with resource consumption, how new services are accessed, and systems in place for risk mitigation, including data and privacy policies and processes. Rather than a cyclically executed process, risk management must be continuous and contain various coordinated actions and tasks in order to oversee and manage risks. An ecosystem-wide framework going beyond traditional IT is necessary for proper risk management. Enterprises must therefore prioritize training and awareness within their organization, teaching team members how to securely use multiple cloud services. 



Quote for the day:

"Integrity is the soul of leadership! Trust is the engine of leadership!" -- Amine A. Ayad

No comments:

Post a Comment