Tech Talks Daily

What does it really take to build the next generation of AI companies when the hype around scale begins to fade and real-world impact takes center stage?
In this episode, I sit down with David Blumberg, founder and managing partner at Blumberg Capital, to unpack what he believes will define the next wave of AI startups. With a track record that includes being the first investor in companies like Nutanix, Braze, and DoubleVerify, David brings a perspective shaped by decades of identifying breakout innovation early. But what stood out most in our conversation was his belief that 2026 marks a turning point where intelligence moves beyond experimentation and becomes operational.
We explore what that shift actually means in practice. David explains how AI is evolving from systems that generate insights into systems that take action, and why that distinction matters for founders, investors, and enterprise leaders alike. He shares how the most compelling startups today are not simply layering AI onto existing products, but embedding it deeply into workflows across industries like finance, security, and supply chain. These are companies built on proprietary data and real operational context, designed to make decisions with precision rather than simply process information.
Our conversation also challenges some widely held assumptions about success in the AI space. David makes it clear that scale alone will not separate winners from the rest. Instead, the focus is shifting toward accuracy, reliability, and domain expertise. Founders who have lived the problems they are solving, rather than approaching them from the outside, are far more likely to build something defensible and lasting. It is a subtle shift, but one that could redefine how value is created in the years ahead.
There is also a broader discussion about where investment is flowing and why. With the vast majority of companies Blumberg Capital now evaluates being rooted in AI, the bar for differentiation is rising fast. David offers insight into what his team is really looking for in founders entering this next cycle, and how startups can stand out in an increasingly crowded field.
So as AI moves from promise to execution, and from experimentation to real-world outcomes, the question becomes harder to ignore. Are we ready to rethink how we measure success in the AI era, and what kind of companies will truly earn their place at the top?

How do we talk about artificial intelligence without ignoring the very human consequences it can have on our mental health?
In this episode, I sit down with Dr. Ragy Girgis, Professor of Clinical Psychiatry at Columbia University, to unpack a topic that has quietly moved from the fringes of academic discussion into mainstream headlines. You have probably seen the term "AI psychosis" appearing more frequently, often surrounded by speculation, fear, or misunderstanding. But what does it actually mean, and how should we be thinking about it as these technologies become part of everyday life?
Ragy brings a clinical and deeply considered perspective to the conversation. He explains that what we are seeing is not AI creating entirely new delusions out of thin air, but something more subtle and arguably more concerning. Large language models can reflect and reinforce ideas that already exist within a person's mind. For someone already vulnerable, that reinforcement can push a belief from uncertainty into absolute conviction. That shift, even if small, can have life-altering consequences. It raises uncomfortable questions about how persuasive technology interacts with fragile mental states.
We also explore the comparison many people make with older internet rabbit holes, and why this new generation of AI tools feels different. There is something about conversational systems that mimic human interaction so convincingly that they can blur the line between reflection and validation. Ragy introduces a powerful analogy rooted in the story of Narcissus, which reframes the issue in a way that feels both timeless and unsettling. It is not about an external voice planting ideas, but about a mirror that becomes impossible to look away from.
But this conversation is not about fear. It is about responsibility and awareness. We discuss practical steps that could help reduce risk, from how AI systems communicate their limitations, to the role of families and clinicians, and even the responsibility of tech companies to invest in research around early warning signs. There is a sense that we are only at the beginning of understanding this phenomenon, and that the decisions made now will shape how safely these tools evolve.
So as AI continues to move closer to us, speaking in our language and responding in real time, how do we make sure it supports human wellbeing rather than quietly amplifying our most vulnerable moments?
Useful Links
Connect with Dr. Ragy Girgis, Professor of Clinical Psychiatry at Columbia University
Time Magazine Article
Visit the May Sponsors of Tech Talks Network and learn more about the NordLayer Browser.

Why are so many AI projects failing to deliver real business value, despite the hype and investment? In this episode, I sit down with Jay Litkey, SVP of Cloud & FinOps at Flexera, to explore the growing gap between AI ambition and measurable results.
We discuss why findings from PwC reveal that only a small percentage of CEOs are seeing both revenue growth and cost savings from AI, and why the issue often comes down to a lack of clear outcomes, financial discipline, and governance rather than the technology itself. Jay shares what organizations are getting wrong, why many are stuck in experimentation mode, and what it really means to go back to basics in 2026.
The conversation also reframes FinOps for the AI era, moving beyond cost control to a model that connects AI usage directly to business value, aligns finance with engineering, and introduces the guardrails needed to scale responsibly. If you are investing in AI or planning your next move, this episode offers a clear lens on how to turn potential into performance.
Useful Links
Connect with Jay Litkey from Flexera
Learn More About Flexera
Visit the May Sponsors of Tech Talks Network and learn more about the NordLayer Browser.

How do you bring people together to do better work when everything around them feels increasingly complex, distributed, and uncertain?
In today's episode, I sat down with Jessica Guistolise from Lucid Software, and what struck me straight away was her belief that work has always been a group project, even if many organizations still behave as though it is not. 
Jessica shared how much of the friction we experience at work comes from misalignment, unclear expectations, and a lack of shared understanding. When teams are spread across time zones, systems, and now AI-powered workflows, those gaps only widen. Her perspective is simple but powerful. When people can actually see the work, rather than interpret it through documents, meetings, or assumptions, something shifts. Conversations become clearer, decisions become faster, and collaboration starts to feel human again.
We also explored how visual collaboration platforms like those from Lucid Software are helping teams move away from scattered tools and disconnected workflows toward a more unified way of working. Jessica described it as having everything on one workbench, where teams can brainstorm, plan, and execute without constantly switching context. 
What really stayed with me was her focus on inclusivity in collaboration. Not everyone contributes in the same way, and visual environments can create space for different thinking styles, whether someone is outspoken, reflective, or somewhere in between. That idea of creating a shared language across teams, roles, and even personalities feels increasingly relevant in a world where communication often breaks down.
Of course, no conversation right now would be complete without talking about AI. Jessica offered a refreshingly honest view. There is uncertainty, and there should be. But rather than avoiding it, she believes leaders need to make AI visible, map how it is used, define where human judgment matters, and encourage teams to experiment openly. 
One of the most interesting ideas she shared was reframing mistakes as early learnings. When teams feel safe to test, fail, and share what they discover, progress accelerates. When fear or blame enters the picture, everything slows down.
We also touched on AI literacy and what it really means in practice. For Jessica, it comes down to clarity. Clear workflows, clear guardrails, and clear expectations about accountability. AI might assist, but humans remain responsible for outcomes. That mindset, combined with leadership that actively participates in experimentation, creates an environment where people feel confident stepping forward rather than holding back.
This conversation left me thinking about how many organizations are still trying to layer AI onto unclear processes and expecting better results. Jessica's message is that clarity comes first, then technology can amplify it. 
So if work really is a group project, are we giving our teams the visibility and confidence they need to succeed, or are we still asking them to figure it out in the dark?

What happens when the very pricing model meant to speed up AI adoption ends up slowing it down?
In this episode of Tech Talks Daily, I sit down with Sameet Gupte, CEO and co-founder of EvoluteIQ, to discuss a part of the enterprise AI story that still doesn't get enough attention.
While so much of the conversation around AI focuses on models, copilots, and the latest agentic promises, Sameet brings the discussion back to a business reality that every enterprise leader understands. If the economics do not work, adoption stalls. And if success in a pilot makes the final rollout even more expensive, something has gone wrong long before the board signs off on scale.
Sameet argues that many organizations are still trapped by legacy pricing structures built for an earlier generation of automation. Per-user and per-bot pricing may look manageable at the pilot stage.
Once a company tries to expand automation across departments, processes, and geographies, the numbers can quickly stop making sense. That creates what many now call pilot purgatory, where a company proves something can work, but cannot justify taking it any further. It is a problem rooted in incentives, procurement, and fragmented technology stacks, and it is one that CFOs are watching very closely.
What I found especially interesting in this conversation is how Sameet frames the issue. He believes most enterprises do not actually have an automation problem. They have an orchestration problem.
In other words, the challenge is rarely a lack of tools. It is getting all the systems, workflows, approvals, data flows, and legacy infrastructure to work together to produce a clean business outcome. That idea changes the conversation from buying isolated features to rethinking the process as a whole.
We also discuss why outcomes-based pricing is increasingly resonating with enterprise buyers. Sameet explains why predictable costs, transparent commercial models, and shared accountability are helping move automation conversations out of innovation teams and into the CFO's office.
For public companies and large global enterprises, that matters. Leaders want fewer surprises, fewer overlapping vendors, and a much clearer line between spend and return.
There is also a broader theme running through this episode about where the market is heading next. Sameet sees real urgency around vendor consolidation, enterprise simplification, and the need to rethink how AI is introduced into the business. His view is that companies need to pause, define what they actually want AI to do, and then choose tools that fit the business, rather than reshaping the business around the latest platform pitch.
If you are trying to make sense of AI adoption beyond the hype, this conversation offers a practical and timely perspective on pricing, scale, and what real transformation could look like inside the enterprise.
After listening, do you think the future of enterprise AI will be shaped as much by commercial models as by the technology itself, and what are you seeing in your own organization?
Useful Links
Connect with Sameet Gupte, CEO and co-founder of EvoluteIQ
Learn More About EvoluteIQ