Federal Tech Podcast: for innovators, entrepreneurs, and CEOs who want to increase reach and improve brand awareness

When the history of software development is written, they will look back at the decade that preceded 2023 as the era where developers started to pull pre-written code off repositories and assemble code in this “modular” manner. Many federal systems are constructed in this “cut and paste” method. This is all well and good if the code can be trusted.  But can it? Today, Emile Monette from Synopsys shares with listeners some fascinating statistics that document this historical transition and he offers solutions as well. During the interview, Emile Monette references a recent study by his company, Synopsys, called the Open-Source Security and Risk Analysis Report. They examined 1,700 codebases across seventeen industries and their findings correlate to what technology leaders think is happening. Ninety-six percent of scanned codebases contained open-source code. Perhaps the federal government should consider ways to make sure this code is safe to use. However, there is even more surprising data that was released. Emile notes that, in addition to examining codebases more carefully, systems administrators should consider the basic “blocking and tackling” of software.  He cites the figure that 89% of codebases were out of date and 91% contained components that were not the current version. Synopsys has been in business since 1986 and has decades of experience in developing code for highly complex silicon design.  This is a rigorous development that must have high assurance of security and quality. They are applying this in-depth knowledge to help the proliferation of federal organizations that are in a situation where they require to know that the code they are using is safe.   Follow John Gilroy on Twitter @RayGilray Follow John Gilroy on LinkedIn  https://www.linkedin.com/in/john-gilroy/ Listen to past episodes of Federal Tech Podcast  www.federaltechpodcast.com

In this podcast, the Vice President of NetApp discusses managing the federal hybrid cloud and optimizing presence in the cloud. They explore NetApp's offerings such as 'Cloud Checker' and 'ONTAP' for cost-effective storage. Blue XP is introduced as a tool for managing the entire IT landscape. They also discuss the concept of tiering for cost optimization and the importance of resiliency in managing satellite systems.

In this podcast, Chris Howard from Vectra discusses how their company applies Artificial Intelligence to the world of cybersecurity. They have developed techniques to understand insider threats and use Attack Signal Intelligence to isolate and respond to malicious actors. The podcast also highlights the importance of understanding data in threat detection, continuous improvement in software development, data exfiltration monitoring, and advances in threat detection and response.

Out of all the phrases that technology has generated in the past ten years, I think I like the phrase, “cloud washing” the most.  The concept is that a traditional company would promote itself as always being cloud native.  Kind of like being a Monday morning quarterback. In 2023, the variation on that theme is “Artificial Intelligence washing.”  Now that AI is a trending phrase you can see technology websites suddenly claiming the long history they have with artificial intelligence, or, what may be called, “artificial intelligence washing.” This brings us to today’s interview with Chris Howard from Vectra. The company has been knee-deep in Artificial Intelligence since 2010.  As Chris explains, they have honed that knowledge to be able to apply it to the world of cybersecurity.  During the interview, Chris explains how Vectra deals with an insider threat.  They have focused on techniques to understand how an insider moves within a network.  The overall approach to this is what Vectra calls “Attack Signal Intelligence.” Without divulging proprietary information, Vectra has proven itself to be able to isolate and respond to a variety of malicious actors. Essentially, they have trained AI to think like an attacker and give the appropriate defensive response.    Chris describes a case study where a person was let into the network with appropriate credentials as a test to see Vectra’s capabilities. Merely by watching behavior, the algorithms from Vectra were able to isolate the malicious actor. Realistically, attacks are so prevalent that defense is almost impossible for a human without an automated tool. The safety provided by Vectra doesn’t stop with the installation of the Vectra platform. It has been designed to continuously learn new threats and respond accordingly. Listen to the interview to gain a better understanding of an advanced way to handle cyber threats.   Follow John Gilroy on Twitter @RayGilray Follow John Gilroy on LinkedIn  https://www.linkedin.com/in/john-gilroy/ Listen to past episodes of Federal Tech Podcast  www.federaltechpodcast.com    

Bill Wright, Global Head of Government Affairs at Elastic, discusses the National Cybersecurity Strategy and the challenge of federal privacy legislation. The strategy aims to fill in security gaps in the private sector, shift security liability to software makers, and incentivize basic security built into offerings. Industry groups may resist these changes. Public-private partnerships are crucial for enhancing cyber protections in critical infrastructure sectors.

When the history of technology of the twentieth century is written, one of the giants will probably be Ray Kurzweil. As most listeners know, he designed the first Optical Character Recognition (OCR) machine.  The drudgery and error-inducing process of keying in forms was reduced.  Today’s interview is with Chirs Harr from Hyperscience.  During the interview, he gives listeners an understanding of how OCR has become Intelligent Document Processing.  He argues that the founders of Hyperscience produced innovation that combines expanding OCR’s ability and have it reducing clerical errors, improving performance, and deliver better customer experience.  Not only that but the solution can also be scaled to handle the enormous number of documents. The ability to scale saves taxpayers money. In a recent study conducted last year, there is a report that four agencies process over 800 million documents a year.  This number seems high until you think about the size of your tax return last year. Handling a massive number of documents applies to artificial intelligence.  It may not have occurred to you that a large part of the information that is poured into machine learning is generated with a paper document. Any effort at increasing the accuracy of that data means the results will improve. Follow John Gilroy on Twitter @RayGilray Follow John Gilroy on LinkedIn  https://www.linkedin.com/in/john-gilroy/ Listen to past episodes of Federal Tech Podcast  www.federaltechpodcast.com

If you toss a baseball anywhere inside the Washington, D.C. beltway, you will probably hit a company that develops software. Well, what differentiates them? Size?  Degrees the staff hold.  The number of contracts?  Today, we sit down with a couple of software development experts from Excella to talk about completing more that the bare minimum for a project. They argue, quite cogently, that quality is the characteristic that is the hardest to achieve and has the most lasting impact on federal projects. Two representatives from Excella are in the studio to argue their case.  Doguhan Uluca and Keith Mealo from Excella sat down to discuss everything from shift left to legacy to elite performance. The interview begins with a discussion about FISMA High Impact and FedRAMP.  From there, they provide opinions on memory safe-languages and cloud-native development. They both have extensive experience with highly sensitive software development projects. Some companies look at a project and set up a checklist and methodically go down the list.  When the minimum is completed, the project is complete. Keith Mealo suggests that one must have full-orb comprehension of the system.  When a person examines the consequence of deciding can you get an idea of solving the problem, not just checking a box. During the discussion, the moderator tried to draw a parallel between an automobile getting a State inspection. A mechanic may see   “examine the seat belt” and hit the check box.  A better inspector may ride with the driver to make sure the safety device is in use.  However, the master mechanic will notice a problem with the brakes while looking at the seat belt and not release the vehicle until the entire vehicle is safe. Follow John Gilroy on Twitter @RayGilray Follow John Gilroy on LinkedIn  https://www.linkedin.com/in/john-gilroy/ Listen to past episodes of Federal Tech Podcast  www.federaltechpodcast.com

Artificial Intelligence certainly has piqued the interest of the defense community here in Washington, DC.   The Brookings Institution has a recent study that shows 307 vendors working with the federal government on many aspects of artificial intelligence. Today, we look at a company that has been involved with helping humans leverage information from data for over 25 years. His emphasis is to be a champion of human and AI teams.   During this interview with Danial Serfaty the founder and CEO of Aptima. He suggests that artificial intelligence may not be the panacea everyone seems to be looking for.  He talks about some of the skills needed to develop effective AI. For example, one must be careful during the machine learning aspect of development. However, one can see some applications for AI in the federal government.  One obvious example is diagnosis x-rays.  Today, machines can see issues that may elude diagnosticians. Daniel Serfaty is not saying that AI should take over the diagnosis, however, AI can be a tool that can assist in understanding medical issues. Also, COVID has put an enormous amount of stress on hospitals.  Let’s say we have a surgeon who is going into multiple procedures late in the week.  Wouldn’t it be effective to be able to ask a medical system that has AI built int to it for an opinion? A second example may be the proliferation of space debris.  There is an estimated xxxx in space right now.  We know 30,000 satellites will be launched by 2030 – how can anyone keep track of a myriad of satellites and thousands of pieces of debris? This is a problem that collecting information and making course projections could help ensure safe satellite transmissions.

Today’s interview with Mike Wiegland from Shift5 highlights how commercial technology can assist warfighters on the battlefield.  Mike provides insight into how the commercial industry can help innovate with defense technology. He has a military and technical background and has real-world examples of how operational technology can enable military systems to deter adversaries. The federal government is being attacked relentlessly. Malicious actors are constantly looking in every nook and cranny to get an edge. Even if you dedicated all the estimated 1.3 million DoD employees full-time, it wouldn’t be enough to overcome the attacks. As a result, federal agencies are looking to partner with private companies to solve some persistent problems.  One can define “public-private” partnership in many ways, the most common are efforts at defining standards. However, what happens when a commercial organization has a technology that can assist a federal agency?  The most recent example was when Starlink assisted Ukraine with geo-intelligence.  Ostensibly, the satellite system was a commercial endeavor, but it was used at the tip of the spear. The referenced  mass of electronics circling the earth at 23,000 is a simple, visual example. When one uses commercial products for a federal, or even a military application, it can be called “Dual Use.” A more mundane example may be operational technology or OT. Frequently, these are sensors that monitor industrial or transportation equipment. An example may be monitoring sensors on aircraft. Follow John Gilroy on Twitter @RayGilray Follow John Gilroy on LinkedIn  https://www.linkedin.com/in/john-gilroy/ Listen to past episodes of Federal Tech Podcast  www.federaltechpodcast.com

In the early days of the Internet, companies would design databases that could handle what they perceived as tremendous amounts of information.  Memory was expensive, so those databases were optimized to handle related pieces of information, or relational databases. They relied on a system that used structured data language to pull up information. It got to the point where two or three companies dominated all large systems. They became so ubiquitous that companies like Microsoft would acknowledge the dominance of Oracle by using it for Microsoft customer names.   The past fifteen years have seen memory get cheap, and virtualization has made storage a minor concern.  Combining this with mountains of unstructured data being generated with sensors, phones, and occupational technology, new databases had to evolve.  One way to break through the constraint of a relational database is with a way to approach that that goes beyond Structured Query Language, or “Not only Structured Query Language.”  MongoDB is an example of a database classified as NoSQL.   Some have classified this as an object-oriented database.  When comparing them, some will argue that a SQL database is better when one has many writes; and a No SQL is faster when it comes to reads.  Like anything, it is best to carefully evaluate your needs and select a system that fits your set of requirements. MongoDB is an example of this No SQL approach. During the interview, Brent Leech provides an overview of why MongoDB has the performance characteristics that will serve as the answer to many federal data management goals.  The federal government is presented with many data handling challenges – they have vast repositories of structured data as well as daily accumulating unstructured amounts of information. The more you know about databases the better you will be at evaluating appropriate systems. Follow John Gilroy on Twitter  @RayGilray Follow John Gilroy on LinkedIn  https://www.linkedin.com/in/john-gilroy/ Listen to past episodes of Federal Tech Podcast  www.federaltechpodcast.com