Daily Tech Digest - August 19, 2023

Inside the Rise of 'Dark' AI Tools - Scary, But Effective?

This shouldn't be surprising, since building LLMs is an intensive endeavor. "As what WormGPT showed, even with a dedicated team of people, it would take months to develop just one customized language model," Sancho and Ciancaglini said in the report. Once a product launched, service providers would need to fund not just ongoing refinements but also the cloud computing power required to support users' queries. Another challenge for would-be malicious chatbot developers is that widely available legitimate tools can already be put to illicit use. Underground forums abound with posts from users detailing fresh "jailbreaks" for the likes of ChatGPT, designed to evade providers' restrictions, which are designed to prevent the tool from responding to queries about unethical or illegal topics. In his WormGPT signoff earlier this month, Last made the same point, noting that his service was "nothing more than an unrestricted ChatGPT," and that "anyone on the internet can employ a well-known jailbreak technique and achieve the same, if not better, results by using jailbroken versions of ChatGPT."

4 ways simulation training alleviates team burnout

Simulation training boosts confidence because unlike traditional training methods, the learner gains experience over time through true-to-life virtual cyber warfare training and sparring against simulated malicious adversaries that behave like human opponents. By training in the same IT infrastructure they have at their job— complete with networks, servers, and security tools—they improve competencies, judgment skills, and gain “muscle memory” so they feel prepared to respond to a real cyber incident. ... With simulation training, SOC teams learn to identify false positives and high-priority alerts more effectively over time as they become familiar with the types of alerts that end up impacting their organization’s infrastructure. The training can mimic the high volume of alerts they receive during the day and help teams develop effective triage strategies to streamline their response processes. Practicing this in simulation allows teams to experiment on their approach and fine-tune it without fear of making a mistake during operating hours.

A managerial mantra in the age of artificial intelligence

The rise of modern management brought forth professionalism through business schools, advocating ethical standards and fostering professional workplaces globally. Often, this professionalism is rooted in the mastery of managerial principles. These principles are created and taught by a variety of business school professors, and they are developed in close collaboration with executives and leaders. Unfortunately, a lot of these ideas have only been applied sparingly due to practical limitations. These limitations may result from the limited time available for decision-making in the corporate world, the need to manage uncertainties, the lack of data and accurate knowledge of the facts, and occasionally even the ignorance of professional principles. ... Organisational thinkers have traditionally identified that this leads to satisfaction, whereby managers have to be satisfied with the good-enough, not necessarily the best, choice. In other words, constraints on time availability lead a manager to do a limited analysis of the impact of a job candidate on future organisational performance. 

Five Challenges in Implementing AI in Automation

Accuracy and bias are two critical, yet recurring issues in AI that require human supervision. For example, generative AI applications are prone to hallucination, or making up facts based on their training dataset. In the same vein, biased datasets fed into a machine learning model can produce biased results. If a financial services firm is using an AI-driven automated system to accept or reject credit applications, for example, it’s essential to avoid well-documented, systemic biases toward women or people of color that may be contained in the training dataset. As we progress toward AI-driven decision-making, it’s critical for humans to remain in the loop, verifying the results generated by machine learning algorithms to check bias and other forms of inaccuracy. Keeping humans in the loop is a critical step toward re-training algorithms to perform more effectively in a production environment. ... Regulating AI is an ongoing issue globally, and the legal field continues to be shaped by emerging technologies including generative AI. 

Mastering Agile Security: Safeguarding Your Projects in a Fast-Paced World

Just ensuring rapid delivery of the product is not enough. The key to Agile success is to ensure that security is an integral part of the process from the beginning. And since agile is an iterative process, and is all about accommodating changing requirements as and when they arise, security must also be part of this iterative process. Regular security reviews and tests whenever there is a change in the product is the key to delivering a working as well as secure product. ... Agile security is not an impediment to the Agile process; rather, it's an essential component that ensures the final product is robust, resilient, and safeguarded against potential threats. It's not about slowing down development but about integrating security seamlessly into every phase of the project lifecycle. ... At the core of Agile security is the Agile mindset. This mindset emphasizes collaboration, adaptability, and constant improvement. Security is not a one-time event but an ongoing effort that requires the entire team's commitment.

Managing Software Development Team Dynamics from Within

In most cases, the whole team will benefit from trying new tools or services every now and then, just to understand patterns and trends. We know we should always be increasing automation. However, especially with things like JavaScript frameworks, up jumps the New Pusher — too keen to adopt the new when no evidence exists that the gains are worth the disruption cost. Or worse, ignoring the disruption cost entirely. The New Pusher can make the team pine for the road not taken, as opposed to do what they should do, and investigate a little on their own time to see how the team will truly benefit from their shiny find. When thinking about adopting a new tool or service the team should not trial it somewhere inconsequential, as that will be neither conclusive nor beneficial. A short examination or study period should lead to a yes/no decision and the use of the tool or platform somewhere of value. Once the pattern is set, the New Pusher can work to that template. The suspicion that people just want to put new experiences on their CV is a little irrelevant. 

How Generative AI Is Making Data Catalogs Smarter

Sequeda explained how generative AI, which leverages conversational, chat-oriented interfaces to surface results from large language models (LLMs), improves productivity and encourages the adoption of a data catalog. With more traditional data catalogs, administrative tasks require more significant manual interventions, time, and some advanced skills and analysis. Smart catalogs remove these barriers by simplifying and automating some of the administrative workflows. As a result, team members in an organization see faster time to value and find it easier to get started with the catalogs. On the data producers’ end, Sequeda said, “Generative AI automatically enriches metadata around the inputs and provides descriptions and synonyms” in the data catalog, smoothing catalog record creation and upkeep. Also, smart data catalogs give data engineers “code summaries” about catalog queries, reducing the time to do DataOps, including any pipeline malfunctions. Using smart data catalogs, consumers find inspiration when the generative AI suggests alternative queries from previous searches and patterns of results. 

Four Myths About Digital Transformation And How To Debunk Them By Modernizing At The Data Layer

A data fabric architecture is essentially a data mesh with an added “abstraction layer” that virtualizes all data into a centralized platform. The benefit is a single pane of glass for all data, virtualized and contextualized for a broader range of business users to work with. The trade-off is that this sudden visibility can be daunting for DX teams newly tasked with untangling all the previously unseen dependencies, vulnerabilities, governance issues, and compliance or security gaps that suddenly appear. All three approaches remain represented in today’s marketplace for organizations to choose from. And while the calculus for making the choice will vary for each company based on their DX goals and level of technical expertise, a common ingredient to success is to prioritize scalable and repeatable processes through automation and low-code wherever possible. ... Choosing the right underlying data architecture is an ongoing balance of matching the pros and cons of the approach to the specific business and operational needs of the organizations. 

A license to trust: Can you rely on 'open source' companies?

Amanda Brock, OpenUK's CEO, which doesn't have a horse in the IaaS race, appeared disappointed with the company's move. "HashiCorp has always been a true open source company, and what Mitchell Hashimoto and Armon Dadgar achieved from a project never intended to be commercialized has been incredible." Brock then asks, "Taking it to an IPO and seeing Mitchell have the apparent wisdom to step aside and allow a more experienced individual to run HashiCorp – but has that also led to its downfall as an open source company?" Her answer is yes. "The statements about BSL are sadly open-washing. It would be wrong to suggest these two ever intended a bait and switch, but they have indeed switched away from open source. The pressure of enabling their competitors with their innovations – an inevitability of open source – did not align with the need to generate shareholder value." That led her to another, bigger question: How much money is enough? Is a lot of money with others generating a lot of money, too, a reason to stop?" She's left "wondering whether had Mitchell remained CEO, this would have occurred?"

Culture Transformation: What leaders need to know

Fortunately, culture only appears enigmatic: There are practical, tangible, measurable ways leaders can properly manage their culture. And it all starts with alignment. Executives need to be on the same page with their leadership teams -- particularly CHROs -- about where their culture stands today and where it’s headed in the future. You might be thinking: “We’re already aligned about our culture.” But it’s not enough to be generally on the same page. The best leaders are synchronised on specific, seemingly small details about their culture and how they affect performance. In one of our client organisations, the goal of being a high-performance culture is behind all decisions. Every leadership meeting keeps high-performance front and centre in their conversation. For instance, leaders might be on the same page about the core values and beliefs -- such as customer-centricity or excellence in safety outcomes -- that they want their culture to embody. But the best path to excellence varies tremendously by industry, market segment, product and more. 

Quote for the day:

"Success is not a random act. It arises out of a predictable and powerful set of circumstances and opportunities." -- Malcolm Gladwell

No comments:

Post a Comment