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

Artificial Intelligence is at the phase where federal and commercial technology leaders are amazed by novel model architectures. Each one can perform a function like data preprocessing, feature extraction, or prediction. The demonstrations are impressive; today we look at the data behind the dazzle. Adam Kowalski from proximal has spent years helping organizations follow a data-centric AI approach. During the interview, he expands on having high-quality, relevant data has precedence over any model architecture. One weakness of generative AI systems is they can be subject to bias. You may have insufficient data. Taking a data-centric approach can eliminate many of the errors we are seeing in AI results. The results of varying from this precept can be disastrous as well as entertaining. In the commercial world, this can result in millions of dollars of loss. In the federal government, there can be much more grave consequences of ignoring the data. It is not all serious. Adam Kowalski mentions an entertaining term from the AI community. If you ask a public language model a question about ancient Rome, you may discover a key player was Abraham Lincoln. This is called a “hallucination.”    = = = Want to leverage you next podcast appearance? www.podscorecard.com Connect to John Gilroy on LinkedIn   https://www.linkedin.com/in/john-gilroy/ Want to listen to other episodes? www.Federaltechpodcast.com  

Cheap storage and fast Internet have made data collection cheap and fast. It is no wonder that we have federal agencies that are drowning in data. Some will argue that the federal government is the largest collector of data in the world. That may be a fantastic fun fact that will dazzle your neighbors, but how is that data being used to move forward the goals of the agency? Today, we sat down with Jason Green from Virtru. He brings his considerable experience to bear when he describes how agencies can optimize that data. He provides his thoughts on digital strategy, collaboration, and safeguarding that data. STRATEGY During the interview, Jason makes a sage observation. He remarks that when looking at the overview of digital transformation strategy, the focus has been on applications and not the data itself. AI will be making increasing inroads around decision-making and serious decisions cannot be derived from dirty data. COLLOBORAGE – When a data analyst is limited to one data set their conclusions can be biased. The need for collaboration is obvious. The “how” is the tricky part. Jason suggests that more careful attention be paid to tagging data so that some fields can be shared and some not. SAFEGUARDING – Years ago, data was guarded with hashing techniques. In today’s fast-changing world, mere hashing is not sufficient. Data stores can be injected with poison data; this means continuous monitoring must be stretched into the data stores themselves. If you would like to learn more about managing data, you may want to attend “DMV Rising” a networking event at Virtru headquarters in Washington DC on September 5, 2024. You will be able to sit in on sessions as well as ask face-to-face questions for subject matter experts. = = = www.podscorecard.com Connect to John Gilroy on LinkedIn https://www.linkedin.com/in/john-gilroy/ Got goin’ to Mars on your bucket list? Listen to Constellations Podcast https://www.kratosdefense.com/constellations/podcasts Want to listen to other episodes? www.Federaltechpodcast.com    

In 2021 the federal leaders recognized a problem with citizen experience on federal websites. As a result, Executive Order 14058 was released dedicated to improving citizen services for federal technology. However, a 2023 McKinsey & Company report states the average score for federal services is only 31%. Justin Fessler from Yext appears on the Federal Tech Podcast to offer some insights on how to improve the citizen experience, especially as it pertains to search. Everyone has laughed at the misleading results that a search engine gives. Everyone has opinions. It is one thing when you search for some obscure historical fact, and another when you are looking for a tax ruling. During the interview, Justin Fessler observes that much of the information citizens require may be sitting in a data store somewhere in the agency. He suggests that generative AI can be applied to these data sets to provide answers more easily and in the context of the user. One concern that is expressed by everyone is the “black box” nature of generative AI. How can you trust the answer from a faceless system? Justin Fessler provides the answer: citations. In the Yext search platform, stacks of data are searched quickly, and citations are offered for the answer. Improving citizen experience means documents can be retrieved for important ranging from legal to emergency services. = = = www.podscorecard.com Connect to John Gilroy on LinkedIn https://www.linkedin.com/in/john-gilroy/ Got goin’ to Mars on your bucket list? Listen to Constellations Podcast https://www.kratosdefense.com/constellations/podcasts Want to listen to other episodes? www.Federaltechpodcast.com    

The Federal Trade Commission recently estimated that nationwide fraud tops ten billion dollars. Most of the efforts to stop this abuse start with identification. However, the federal response has been weak.  Today, we have a former federal leader and current Vice President and Head of Public Sector Strategy for Society on the podcast, Jordan Burris. He opens with the obvious – if malicious actors are leveraging artificial intelligence to attack identities, then the only logical response is to use AI against AI. AI allows malicious actors to assemble identity information for fast access. Recent reports highlight a concept called an “identity access broker.”  This is a company that stores compromised credentials and sells them like one would buy shoes. From his perspective, the federal government has not moved away from legacy systems fast enough and is leaving itself vulnerable. For example, a traditional system may approve an identity, but not be able to continuously maintain that identification. It has been shown that Multi-factor Authentication can leave sessions open where an initial valid identity can be compromised. Identity is not stagnant and innovative approaches can look at many aspects of identity, including liveness detection. CrowdStrike’s Global Threat Report shows a 147% increase in identity attacks. Federal agencies are not immune to these attacks.  = = www.podscorecard.com Connect to John Gilroy on LinkedIn https://www.linkedin.com/in/john-gilroy/ Got goin’ to Mars on your bucket list? Listen to Constellations Podcast https://www.kratosdefense.com/constellations/podcasts Want to listen to other episodes? www.Federaltechpodcast.com  

ChatGPT has overwhelmed the headlines in the past few months. It has brought the concept of artificial intelligence to the forefront of many discussions. Like everybody talking about swimming in the Olympics and then forgetting about it for four years. Well, Primer.ai has been deeply involved in artificial intelligence since its inception seven years ago. It did not start with creating a bot for a site selling shoes, it got its start from I-Q-Tel and has a laser focus on helping the defense community apply technology to real-world problems. The company has deep relationships in the defense community and understands the need of applying natural language processing to the warfighter’s needs. During today’s interview, Mark Bruner from Primer set up the discussion by talking about innovation from Silicon Valley and the requirements of the hardnosed military community. Rather than walking away from the potential benefits of technical innovation, Primer sees itself as an intermediary with the skill set to be able to bridge the gap between real-world requirements and the software community. One of the key elements that is glossed over is the need for ethical application of artificial intelligence. This means understanding how large libraries are assembled and then, at the output end, knowing when to assist the data analysis with human values. The military knows it must adapt to a rapidly changing world, Primer can broker relationships and see value where others cannot. Here is an article from Primer that may help  To go faster on AI, Start with Existing Gaps https://primer.ai/featured/to-go-faster-on-ai-think-small-and-build-trust/ = = = www.podscorecard.com Connect to John Gilroy on LinkedIn https://www.linkedin.com/in/john-gilroy/ Got goin’ to Mars on your bucket list? Listen to Constellations Podcast https://www.kratosdefense.com/constellations/podcasts Want to listen to other episodes? www.Federaltechpodcast.com              

In most of my interviews, I sit down with a company and talk about innovation in solving federal technology goals. Today, a twist. Sonny Hashmi was at the center of the battle for improving federal technology. For the past three years, he was the Commissioner of the Federal Acquisition Service at the GSA. This has given him a unique perspective on what the bottlenecks are in streamlining federal technology. He dealt face-to-face with issues like Zero Trust, remote work, and digital innovation. After looking at a myriad of issues, he decided that the way to apply Occam’s law to federal technology was to reduce the complexity of software development. Sonny Hashmi explains that 80% of the technology budget is spent on maintaining systems. This means continuous monitoring, patching, and applying new compliance regulations. His approach to solving the problem is to start at the beginning, software development itself. The solution is quite simple:  adopt a codeless method that allows a wider range of people to work with legacy systems, rapid application, and compliance. During the interview, Sonny Hashmi details some problems he has seen. He talks about everything from struggles with ATO to end-of-life application management. Listen to the interview to learn how the Unqork approach can assist in many aspects of the major digital transformation the federal government is undergoing. www.podscorecard.com Connect to John Gilroy on LinkedIn https://www.linkedin.com/in/john-gilroy/ Got goin’ to Mars on your bucket list? Listen to Constellations Podcast https://www.kratosdefense.com/constellations/podcasts Want to listen to other episodes? www.Federaltechpodcast.com            

Everyone reading this has seen the movie Top Gun with Tom Cruise and heard the classic phrase, “The Need for Speed.”  Well, a desire for speed is not limited to Naval Aviators. People in federal technology may not have call signs but need to optimize communication speed between data centers. If you were to stand in front of a whiteboard with a bunch of enterprise architects, they would dazzle you with CPU and GPU speeds. Memory tricks with virtualization give you tons of memory. However, at the end of the day, the real choking point is the connections between those stacks of silicon. During today’s interview, Rob Shore from Infinera unpacked how optical speeds have drastically changed over the years. Today’s speeds allow for transmittal of 1.2 Terabits per second. That makes the speed of your home wi-fi look like a bicycle racing a Maserati. The conversation took a fascinating twist. While everybody is debating the methods of data collection for artificial intelligence, systems engineers are worrying about the hardware being able to “catch up” to the speed demands that, for example, driverless cars demand. At the end of the conversation, Rob talks about applying the concepts for fast optical cable to inside the data center. Same principles, a need for speed between servers and routers inside a data center as well as between data centers. You can watch the short video starring Rob Shore that explains one technical aspect of technology, coherent optics. = = = = = =  Got a podcast interview coming up?  What's your score?  www.podscorecard.com Connect to John Gilroy on LinkedIn https://www.linkedin.com/in/john-gilroy/ Got goin’ to Mars on your bucket list? Listen to Constellations Podcast https://www.kratosdefense.com/constellations/podcasts Want to listen to other episodes? www.Federaltechpodcast.com            

An American poet once wrote about reaching a fork in the road and he considered which path to take.    When one attempts to consider branching logic and automation in complex technical systems, there is a lot of consideration given to which path to take. There may be lessons to be learned from Robert Frost’s 25-line masterpiece.     Today’s discussion is with Thomas Kinsella, Chief Customer Officer and Co-founder of a company that focuses on improving automation.  It is appropriately called Tines.  Thomas Kinsella was asked to discuss a typical day in the life of a person working in a Security Operations Center, or SOC.  He describes it as a plethora of alerts, new data, and disjointed tools.  His description of the SOC was kind of like One Day in the Life of Ivan Denisovich by Aleksandr Solzhenitsyn. The story is almost as monotonous as a Soviet era gulag, although the remuneration may be slightly higher. Thomas Kinsella describes issues with alert fatigue, difficulty of finding good staff, and incompatible systems that make duplicative works. He and his co-founder decided to come up with a better solution, they founded Tines. During the interview Thomas Kinsella describes how they have been able to use technology to assemble data from a variety of sources. This was to address the idea of alert fatigue, if an incident was possibly confirmed, then it should be pursued. Secondly, the way to address staff shortage was to design a visually based system so that an untrained individual could drag and drop actions into a sequence.  In fact, Tines provides playbooks for many scenarios in a secure environment, he calls them playbooks.  The result is a system that can produce a Secure Orchestration Automation and Response system that is easy to use, scalable, and doesn’t need the expertise of a PhD. in computer science. For more on Tines, please read the blog “Cybersecurity in 2024: Five Predictions from our co-founders”  = = = What’s your score? Take the Podcast Appearance Scorecard www.podscorecard.com Connect to John Gilroy on LinkedIn https://www.linkedin.com/in/john-gilroy/ Got goin’ to Mars on your bucket list? Listen to Constellations Podcast https://www.kratosdefense.com/constellations/podcasts Want to listen to other technology podcasts? www.Federaltechpodcast.com              

Today we sit down with Susan Kidd and Valinder Mangat and discuss FedRAMP. Some will argue that FedRAMP is an unruly thirteen-year-old. Although FedRAMP has authorized over three hundred public service cloud providers, there is much work to be done. For example Valinder Mangat opens the discussion by noting that we have approximately five hundred applications in the pipeline. We have a situation where manually reviewing compliance is going to be too difficult. As a result, FedRAMP is making changes. They are starting to automate the process, causing companies to become familiar with a new acronym, Open Security Controls Assessment Language, or OSCAL. During the interview, Susan Kidd reviews her philosophy that ties into understanding OSCAL. It has been her experience that there is a limit to working hard; the best results are accomplished when one takes advantage of automation and can work smart. To that end, Susan Kidd launched something called Idea Labs, an initiative that can assist federal agencies in modernizing their automation processes using OSCAL. Today’s compliance is not just a check box item. In the interview, Valinder Mangat details aspects of software development. There was a time when code was released and approved, like a snapshot. Today, there is a continuous improvement model in effect along with continuous testing. The only way to accomplish that is to leverage technologies like OSCAL to keep up with changes in threat actors and best practices for handling mountains of data. Valinder Mangat has a nice summary of this consideration:  champions of innovation instead of captives of compliance. In other words, technology can be leveraged so a company does not take six months to get approved. OSCAL puts tools into the hands of federal leaders faster. More details about the IdeaLab at DRTConfidence.    = = = What’s your score? Take the Podcast Appearance Scorecard www.podscorecard.com Connect to John Gilroy on LinkedIn https://www.linkedin.com/in/john-gilroy/ Got goin’ to Mars on your bucket list? Listen to Constellations Podcast https://www.kratosdefense.com/constellations/podcasts Want to listen to other technology podcasts? www.Federaltechpodcast.com              

If you were to take a class in Data Management 101, you would walk into the classroom on the first day and be bombarded with terms like data lake, data warehouse, and data mart. Phew. Where to begin? Let us jump to a federal agency. They are being deluged with data from low code no code, big application platforms, and let us not forget legacy applications that are not in the cloud. How to manage this ball of data confusion. Once it is overseen, how can an agency share information with outsiders to allow for data collaboration? We begin today with a solution offered by a company called Snowflake. Winston Chang suggests that if can take advantage of a single platform, you can abstract the data layer which allows you to manage the data more effectively. OK. Now that we can look across data sets, we encounter the problem of what eyes get to see. Who gets to see what columns and rows and who does not? Winston Chang suggests that a data platform will enhance collaboration by allowing leaders to be able to decide which information, or parts of information, will be available to which parties. During the interview, Winston makes some provocative statements. He argues that if we want to focus on infrastructure, then the data itself must be considered as the infrastructure. If this discussion provokes more detailed information, you may want to consider attending the “Data for Breakfast” conference on March 7, 2024, at Tyson’s Corner Ritz Carlton. You can question experts from  Snowflake on topics as varied as artificial intelligence and managing data as a platform. What’s your score? Take the Podcast Appearance Scorecard www.podscorecard.com Want to listen to other technology podcasts? www.Federaltechpodcast.com Connect to John Gilroy on LinkedIn https://www.linkedin.com/in/john-gilroy/ Got goin’ to Mars on your bucket list? Listen to Constellations Podcast https://www.kratosdefense.com/constellations/podcasts