Invisible Technologies on RLHF and LLM Training

“Source code became almost free, like cheap and not important. And binaries became king because this is the outcome,” says Shlomi Ben Haim, co-founder and CEO of JFrog. Ben Haim joins Bloomberg Intelligence analyst Mandeep Singh in this episode of the Tech Disruptors podcast to discuss the impact of large-language-model coding agents, the challenges of keeping up with model guardrails, and why software supply-chain security and governance are top of mind for modern enterprises. Ben Haim explains how JFrog acts as the infrastructure and control plane for developer workflows, using JFrog Boost, a Model Context Protocol registry and JFrog Curation to help customers optimize token consumption, block malicious packages and automate software governance without compromising speed.

“The minute it worked, we actually started to use smaller models for the things we knew that it worked and see what they could accomplish at a fraction of the price,” says Matt Hicks, CEO of Red Hat, in a discussion with Bloomberg Intelligence Senior Technology Analyst Anurag Rana. In this episode of the Tech Disruptors podcast, the pair discuss Red Hat’s role inside IBM, the durability of hybrid cloud and why OpenShift, virtualization and AI are becoming key growth drivers. Hicks explains how Red Hat is applying AI across engineering and business operations, using smaller models, containers and OpenShift to help customers build more flexible, cost-efficient AI infrastructure.

“Data centers are here to stay, and the world is going to need them as AI adoption accelerates,” says Steven Siesser, partner at Lowenstein Sandler, chair of the firm’s private equity practice and co-lead of its data center practice, in conversation with Bloomberg Intelligence’s Associate Director of Research Alexandra Davidov and Senior Technology Analyst Anurag Rana. In this episode of Tech Disruptors, they discuss how AI data centers are reshaping infrastructure finance, from hyperscaler-backed leases and neo-cloud credit risk to private equity, private credit and sovereign capital. Siesser explains why power, community acceptance and supply-chain constraints remain key bottlenecks, while long-term leases and hyperscaler demand are reshaping how lenders underwrite the build-out.

The cybersecurity industry is facing unprecedented challenges as AI models become more capable. On this episode of Tech Disruptors, SentinelOne co-founder and CEO Tomer Weingarten talks about deploying security for AI agents and the evolution of the security operations center (SOC) with LLMs. He joins Bloomberg Intelligence analyst Mandeep Singh to discuss the impact of Anthropic Mythos, M&A in cybersecurity and the changing nature of attacks as tools become more sophisticated.

Matt Fitzpatrick, CEO of Invisible Technologies, joins Bloomberg Intelligence’s Mandeep Singh on this episode of the Tech Disruptors podcast to discuss the use of reinforcement learning by frontier model providers for training, as well as the company’s enterprise business. They explore reinforcement learning from human feedback (RLHF), agentic AI and self-improvement, the evolution of large language models, coding agents and contact centers.

Artificial intelligence is reshaping banking infrastructure, from accelerating core migrations to enabling agentic experiences that can automate complex workflows. In this episode of Bloomberg Intelligence’s Tech Disruptors podcast, Vishal Dalal, chief product officer of financial solutions at Fiserv, joins BI fintech and payments analyst Diksha Gera to discuss how banks are using AI to modernize legacy systems and improve data quality while managing security, compliance and governance risks. The conversation explores Fiserv’s vision for AgentOS, an intelligence layer that sits on top of bank data and applications, enabling agents to interact dynamically across platforms.

As AI infrastructure scales up, the conversation is moving beyond graphic processing units (GPUs). Faster compute creates new pressure on the data layer, and companies are increasingly focused on whether their infrastructure can move, manage, protect and deliver data fast enough to keep AI systems productive. In this episode of Tech Disruptors, Bloomberg Intelligence analyst Woo Jin Ho speaks with Alex Bouzari, CEO of DDN, about the company’s role in AI data infrastructure, the shift from storage to broader data platforms, DDN’s high-performance computing heritage, its work with Nvidia and the business-model changes taking shape as AI moves from experimentation to production.

“It’s clear to us that the world is short compute right now, and the industry is racing to catch up,” Rob Bittencourt, partner and head of thematic investing at Apollo, tells Bloomberg Intelligence’s Alexandra Davidov and Paul Gulberg on this episode of the Tech Disruptors podcast. Bittencourt discusses why AI is becoming a private-credit story, with trillions of dollars of data-center, power, chip and infrastructure investment needed to support the next phase of adoption. He also explains how Apollo underwrites AI infrastructure risk, why hyperscaler demand and investment-grade financing matter, and how investors should separate temporary software valuation resets from true business-model disruption.

“Every data center will need some part of AI capabilities to run workloads like inference, because it’s just becoming such a fundamental part of all of this cloud technology,” says Alistair Speirs, general manager of Microsoft’s Azure Infrastructure. Speirs joins Bloomberg Intelligence Senior Technology Analyst Anurag Rana on this episode of the Tech Disruptors podcast to discuss how AI infrastructure is reshaping the modern data center, from liquid cooling and dense networking to custom silicon and distributed supercomputing. Speirs explains how Azure is preparing for a world where training, inference and traditional workloads increasingly converge, making software-defined infrastructure, power availability and global scale central to Microsoft’s cloud strategy.

“This year is a much more radical technology shift and the most consequential, I believe, ever. But still, you need all foundations of the house to be in order,” Sebastian Steinhaeuser, Chief Operating Officer at SAP, tells Bloomberg Intelligence Senior Technology Analyst Anurag Rana. In this episode of Tech Disruptors, the pair discuss SAP’s autonomous enterprise vision, the rise of Joule assistants and agents, and why AI may strengthen the case for cloud migration, data modernization, and application consolidation. Steinhaeuser explains how SAP is embedding business process context, governance, and industry-specific knowledge into its agentic layer while navigating shifts in software pricing, model strategy, and customer demand for measurable AI adoption.