One of the most thought-provoking mainstage speakers at our Impact 2025 conference last month was Rana el Kaliouby, an investor, entrepreneur, and author who is also host of the podcast Pioneers of AI. Her autobiography, Girl Decoded, was published in 2020.
In our conversation, we talked about AI agents; how vibe coding is being used in large organizations; the prognostic accuracy of the movie “Her”; and el Kaliouby’s biggest concern about the AI age.
Below is a lightly-edited transcript.
• • •
What are you trying to invest in at Blue Tulip Ventures? What’s your thesis about where AI is going?
I’ve been in the AI space forever, but I do think this is a really amazing moment for AI. Our thesis is human-centric AI, so AI that amplifies and augments human ability. We do not believe that AI should take over humanity…. We really think it can help unlock human potential. And so our fund thesis is to look for opportunities in three verticals. One is AI as it applies to health and wellness — anything that can revolutionize and make health care…accessible to everyone. Our second is what we call productivity and learning — future work, financial inclusion, education, and where AI can really reimagine that whole space. And the third bucket is sustainable living — anything related to advanced manufacturing… physical AI, robotics, climate, energy, food innovation. It’s a big mandate, but we have a…very specific human-centric rubric. Of course, ethics is near and dear to my heart.
Can you talk about that? Do you feel like everybody thinks about this through an ethical lens and a human-centric lens? Or are there companies out there that are just like, “Hey, let’s go as fast as we can. AGI is the goal — artificial general intelligence” — and they don’t care about that other stuff?”
…One of the key criteria [is] are [startup founders] thinking intentionally about building this in a responsible way? That can be anything from how they develop technology — are they thinking about algorithm and data bias, for example? Are they really thinking about the use cases and potential unintended consequences, and trying to sort through that?
I would love to hear a couple examples of companies that you’ve invested in that fit into those three buckets?
There’s a company called SynthPop. I actually invested in this company before we started Blue Tulip. …They’re building an AI healthcare administrator. …The administrative cost of healthcare is just insane. And it’s not a great use of peoples’ time. They built this voice agent modeled around the CEO’s voice. The agent does patient intake. It actually calls on behalf of the patient to figure out your insurance codes and whatnot.
Does the agent divulge that it’s an agent when it’s calling my insurance company?
It does not. Obviously, it sounds human, because it’s modeled around [the CEO’s voice.] But it does things like it says “like,” or it says “umm.” …[They] got a call from the CEO of Blue Cross Blue Shield, because it had identified this call number that was making hundreds of thousands of calls. Eventually we’ll get to the world where agentic AIs are talking to each other.
Any other examples that come to mind, particularly for an audience of people in larger, more established companies?
One of the companies we’ve invested in is called Tough Day, and they are building AI managers. The idea is that your good managers don’t scale. And also, there’s a lot of bad managers out there, and all of us have had tough days at work, right? The idea is they can go to Tuffy, this conversational AI agent. Maybe they have really simple questions around HR policy, health insurance. But sometimes you have really sticky situations, and you don’t necessarily want to escalate to human resources, but you also maybe do not have a network of peers… Our first use case will be the Employee Association of Hawaii. This organization works with thousands of companies based in Hawaii, and so they deploy that agent across the entire network of companies. The average usage time is about 45 minutes. There’s a lot of conversation going back and forth. …They did have to train the agents, though, to have an aloha spirit, because this company is based in New York…
Typically, you go to the engineering team and your idea would get added to the backlog, and it probably would never see the light of day. But with vibe coding, you can just build your idea…
What if you’re one of these companies out here — a Caterpillar, a USAA, a General Mills type company. What should you be doing? How should you be seeing AI as an opportunity right now?
I recently interviewed Anton Osika, who is the CEO of [the “vibe coding” platform] Lovable. It’s one of the fastest-growing vibe coding platforms. I initially thought that most of the use cases [would be] these solo founders who don’t have a technical background, who vibe code the initial version of their product. There’s definitely a lot of that going on. …But the use case I didn’t realize, which was where they’ve seen a lot of growth, is companies where [you may be in] the accounting department or in the marketing department, and you have an idea. Typically, you go to the engineering team and your idea would get added to the backlog, and it probably would never see the light of day. But with vibe coding, you can just build your idea, and you can create these tools. My point is … a lot of these larger organizations are embracing AI to bring a lot of tools, internal tools, to bear in a way that was not possible before. And I think that’s really exciting,
Your podcast, which I mentioned before, [has] had some amazing guests, including Sal Khan from Khan Academy, Vinod Khosla, Mark Benioff. Is there someone you’ve interviewed recently that’s got you thinking in a different way about either the opportunities, or the risks or the dangers, or how AI is going to evolve?
One is Karen Hao, who wrote the book Empire of AI. …The book is amazing, and it’s a must-read.
She does a deep dive into [OpenAI CEO] Sam Altman’s world, and it’s fascinating, because [it explores how] a lot of power is concentrated in the AI space, and that worries me. It’s one of the main reasons I decided to become an investor. I feel so strongly that we need more diverse founders building AI, but also even more diverse investors…
My biggest concern is what [AI] does to human relationships. That’s far my biggest concern. …I think the line is very fine between a copilot that helps you be productive, and an AI companion or partner that takes you away from human-to-human relationships.
You said you worry about the concentration of power. Are there any of the sci-fi scenarios that you worry about, like one day, AI is takes over every data analyst job, or autonomous drivers take over every truck driving job? [What do you think is] plausible enough to be worth thinking about?
My biggest concern is what it does to human relationships. That’s far my biggest concern. You’re seeing a lot of these AI companions, AI friends, AI copilots. I think the line is very fine between a copilot that helps you be productive, and an AI companion or partner that takes you away from human-to-human relationships. I don’t think we’ve figured that out.
I have a 16-year old son who’s very AI-forward, but thank goodness he does not yet have an AI girlfriend or AI best friend. But I worry about that. I think it was Yuval Harari who said that the last decade was about a race for human attention, and the next one will be a race for human intimacy. …It will be about maximizing time spent with the bots, which will take away from our human relationships.
I know — I worry about my Tough Day agent boss. What if I like their advice much better than my human boss? Why do I want to go to that one-on-one, versus chatting with the bot?
My daughter is 22. And apparently, her generation, [has something] called un-bossing. They don’t want to be bosses to anyone. They don’t care about this whole managerial ladder, but they also don’t want to be bossed, either. They want to have managers or be managers, but they’re very comfortable using AI. …I think we will see future organizations look very different.
I do think the future org chart will be a combination of AI coworkers and human coworkers…
[This audience question] ties into that: building human-centered AI systems that aren’t too human, yet are still usable. I keep going back to the movie “Her,” which explored a lot of these ideas…
“Her” is one of my favorite AI-related movies. It’s a decade old, but if you watch it today, it’s exactly how our world is unfolding… Theodore is this really depressed young guy who downloads Samantha, a really emotionally intelligent and smart operating system, with Scarlett Johansson’s voice. But he gets emotionally attached to her. What I love about it is…she has so much emotional and social intelligence, she is actually able to get him out of his depression and, help him see the world in a new light….
There was a pod series [with] about eight episodes called Bot Love, where they featured real humans who have AI partners. Some of them are in actual relationships, but they still have this AI partner, and some of them opted for an AI partner instead. They’re real humans. The conversations are really vulnerable and real, and it’s really fascinating and, in my opinion, scary. …I think it’s incumbent on companies that are building these bots to not just maximize engagement and time on platform… If you are spending more than four hours a day with your [AI manager] Tuffy, maybe it should say, “No, Scott, go talk to your actual boss, or your actual human coworker.” I don’t know.
Another audience question: Technology is building so rapidly. Do you think that consolidation will eventually happen, and large players will end up out-muscling the smaller startups?
The way I think about is that OpenAI and Anthropic and Microsoft are building the infrastructure for this AI world, and there will be a lot of space and opportunity for companies that build on top of that… If you have defensibility, which comes in the form of access to company-specific data, or data that maybe an OpenAI does not have access to, or you’re solving a very specific problem for a very specific customer. I think that’s defensible. I think there will be a whole slew of vertical AI applications or applied AI built on top of these infrastructure layers that would be very vital and exciting.
What are you actually seeing regarding corporate adoption of AI that was built by startups, and not anyone called OpenAI or Microsoft?
Does anybody know of Harvey AI? They’re in the legal space. I had the co-founder and president on my on my podcast a few weeks ago. …They are partnered with a lot of the big legal firms out there. I assume some lawyers use ChatGPT. I think that’s dangerous. But Harvey AI is built specifically for legal use cases, and hopefully with a lot of guardrails around these use cases.
I think we’re going to double down on human connections.
There’s one more question that I’d love to get to… this world of AI generating sales emails that get deflected by AI spam filters. AI content creation. How can humans thrive as this battle roars around us?
I have a thesis around this. I think we’re going to double down on human connections. We’re just about to make an investment in a company that is using AI to build this relationship intelligence that can help you actually get a warm intro to somebody you want to get connected to, or a company you want to get introduced to… because all of this AI noise does not really build trust, and we’re going to have to double down on trust…
I feel like as an investor, you said you have some concerns. But as investor, you have to believe, “Hey, this is going to create some new, important companies, create jobs. Good stuff will happen.”
I think my message to everyone here is we have agency. We’re all founders, business leaders. We’re adopters of AI. We have agency. We get to choose where we think AI gets built and used, and where it shouldn’t be. And I just try to anchor on that agency.
Do you think that there is an AI bubble that we’re in right now, and how does that end? I want to ask you a tough prediction before we wrap up.
It’s an “and” for me. Pieces of this feel like a bubble, but honestly, we’re so early, there’s still such massive opportunity that we haven’t seen unfold yet. So I don’t know what to make of it.
You must be logged in to post a comment.