Ep. 169 - AI’s Impact on Datacenter Power

While most of us have a basic understanding of phishing, the email attack tactic that’s been with us for years, attack techniques and the attacker community have been evolving rapidly. The speed of that evolution is outpacing the ability of users to detect the latest instances of phishing attacks and the advances are not slowing down. Poornima DeBolle, co-founder and chief product officer at Menlo Security, joins host Eric Hanselman to look at how phishing has changed and the options available to defenders. The days of misspelled, badly worded emails are disappearing behind us, being replaced by ever more targeted versions, enhanced with the power of AI. Defenses are being challenged, from new tactics to appear more trustworthy, to advanced proxies that can subvert multifactor authentication. There’s a greater need for technical protections, as attackers get better at outfoxing even the best trained users.   Host: Eric Hanselman Guest: Poornima DeBolle  

IT buying sentiment has remained positive in the latest edition of the 451 Research Tech Demand Indicator study. It showed a strong start to 2024, but there were other signals buried in the data. Liam Eagle and Malav Parekh, two of the study’s authors, return to the podcast to join host Eric Hanselman to explore the results and their implications. The study explores spending sentiment and contrasts it with macroeconomic impacts, looking at shifts across demographics and technology verticals. Information security spending stays strong, but the gap between expectations in consumers is notable. High-income households are much more positive than their lower-income counterparts. The study data can inform those trying to make sense of the IT landscape. Tech Demand Indicator webinar Link to the 451 Alliance application  

Recent study results are identifying where AI is being put to work and how customers and end users are reacting to its use. Sheryl Kingstone, head of the experiences team, returns to look at the Voice of the Connected User Landscape: Connected Customer, Disruptive Experiences 2024 study with host Eric Hanselman. This longitudinal study tracks changes in attitude and use of AI. While there is significant use of AI, there are strong differences between generations across the study. One of the questions the study raises is how users will work with AI capabilities that are integrated into products and services. There has been a slight positive shift in attitudes as there is more practical use in writing assistance and content creation. Enterprises are starting to integrate traditional AI techniques, such as machine learning, with generative AI capabilities and having success in customer service applications. It’s a complex landscape that impacts how AI will be viewed and leveraged.

Sentiment analysis is a common use case for large language models in generative AI and we’re kicking off Pride Month with a discussion of a recent study that examined the way firms are talking about LGBTQ+ issues in public statements. Returning guest Emily Jasper and Ilan Attar of Pronto NLP, who led the analysis, join host Eric Hanselman to look at what the study explored, how they trained the model used for the analysis, and some of the results. While many organizations make supporting statements around diversity, equity and inclusion, insights can be gleaned on the depth of the commitment expressed. It requires much more than the typical sentiment analysis, not only because of the common use of the word pride, but also because of the complexity of the environments in which supporting statements are made. In a year where DEI initiatives have been under strain, determining the level of “pinkwashing” that may be taking place can offer useful perspectives on positioning and support.

The power required to fuel the wave of generative AI use could be massive, but there are many factors that will determine when the crest of that demand will arrive. Analysts Leika Kawasaki and Dan Thompson join host Eric Hanselman to look at recent research in current demand profiles and discuss where it’s headed. The demand for AI infrastructure is here today, but additional datacenter capacity takes years to build. Feeding AI hunger has become a supply chain problem. It’s demand that can reach 92 MW and will certainly go higher, when the full force of AI requirements are evaluated.

As organizations push to realize the benefits of AI, they are increasingly challenged to deliver the necessary data to fuel their initiatives. While most have data repositories to draw from, the quality and accessibility of that data can be an issue. Shiv Trisal of Databricks and John Schirripa of S&P Global Market Intelligence, join host Eric Hanselman to explore the nature of these challenges and ways to address them. One of the major benefits of AI tooling is the democratization of access to data and it has to be a primary goal, but it’s not always easy to achieve. It means pulling together complex data sets and being able to wrap the necessary controls around them. Governance concerns demand better data quality at the outset, to make the path to implementation simpler. This week’s guests:Shiv Trisal, global industry lead for manufacturing and industry at DatabricksJohn Schirripa, VP, product management, distribution solutions, at S&P Global Market Intelligence

In this special report from the RSA Conference, we hear from seven of the team who were there at the event. It’s a set of quick takes on the latest trends and take aways from the conference in discussions with host Eric Hanselman. Whether it’s AI, cybersecurity or M&A, there was a lot going on and a lot to sort out. Guests:Scott Crawford, Research director for information securityGarrett Bekker, Principal research analystDan Kennedy, Principal research analystPaige Bartley, Senior research analystMark Ehr, Principal research analystJustin Lam, Senior research analystBrenon Daly, Research director, M&A

The flexibility and adaptability of cloud-based infrastructure brings great value to enterprises, but most have been wrestling with managing cloud costs. Melanie Posey, Jean Atelsek and William Fellows join host Eric Hanselman to look at FinOps, the cost management initiative that is looking to regain control of infrastructure costs in this new model. Hyperscalers are expanding their tooling and a bevy of startups are angling to optimize both operations and deployment. It’s not a simple path.

The NAB show, what’s become not only a venue for broadcast innovations, but also the latest in consumer electronics, has just wrapped up. Analysts Justin Nielson and Peter Leitzinger join host Eric Hanselman to discuss their thoughts on the show and perspectives on NextGen TV, AI capabilities and advertising. For over 100 years, the NAB show has served as a forum for content alignment and it’s taking on more. High-definition television got another boost with advances in ATSC 3.0 delivery.

The annual gathering of the business side of the cybersecurity industry, the RSA Conference, will kick off in early May. Scott Crawford and Dan Kennedy return to preview what they’ll be expecting with host Eric Hanselman. The upheaval in the security markets is being mirrored at the conference, with one major vendor deciding not to exhibit and consolidation taking place. All of this takes place as security teams struggle with operational scale. AI to the rescue? We’ll have to see.