# What Founders Have To Unlearn To Become Great CEOs

Podcast: Lightcone Podcast
Published: May 10, 2026
Reading time: 22 min
Canonical: https://podbrew.app/briefs/lightcone-podcast-what-founders-have-to-unlearn-to-become-great-ceos

Spenser Skates, the founder of Amplitude, shares his journey of building the company from a YC startup into a public entity over more than a decade. He details the profound personal and organizational reinvention required to navigate this growth, including the significant shift from a hands-on founder to a large-company CEO.

Skates describes Amplitude's evolution, focusing on the team's initial skepticism towards AI and the critical realization that a new wave of analytics demanded a complete overhaul. He recounts the hard reorganizations, bottom-up experiments, and crucial mindset changes that enabled Amplitude to regain its speed and integrate AI effectively. The discussion also covers the fundamental differences between SaaS-native and AI-native development and the future impact of AI on analytics.

This conversation offers invaluable lessons for founders and leaders on adapting to monumental change. It illuminates the practical challenges incumbent tech companies face in embracing disruptive technologies like AI and provides a blueprint for strategic transformation, organizational alignment, and the evolving demands of leadership in a scaling company.

## Key takeaways

- Incumbent tech companies face significant hurdles in reorienting their structure and processes to effectively integrate new technologies like AI.

- External pressure from investors and sales teams can force companies to develop an AI strategy, even when internal skepticism exists due to the technology's nascent capabilities.

- Integrating new technologies like AI into a large organization with existing product priorities can take up to a year, requiring significant internal effort beyond external pressures.

- Structured internal events like 'AI Week,' which combine leadership exposure, training, and live coding demonstrations, effectively educate and align product, engineering, and design teams on new technological capabilities.

- AI product building necessitates a technology-first understanding of model capabilities, as customers may not be aware of what new AI tools can achieve.

- Unlike many new technologies, AI often sees top-down adoption within companies, with management initiating its use rather than engineers pushing for it from below.

- Sam Altman's salesmanship played a pivotal role in creating top-down AI adoption, convincing key stakeholders of its importance before capabilities fully matured.

- Effective AI transformation within companies requires executives and founders to directly engage with the technology to understand its potential and drive implementation.

- Amplitude's "AI Week" successfully ignited bottom-up innovation, leading to the creation of impactful, unplanned products.

- Amplitude replaced SaaS-native leaders and acquired AI-native founders from YC companies to facilitate its AI transformation, indicating a strategic shift in required talent and mindset.

- The core difference between SaaS-native and AI-native development lies in their approach: SaaS-native is customer-feedback driven, while AI-native involves reimagining solutions based on the state of the art in AI.

- AI-native systems often require a different user mentality, involving iterative trial-and-error and a high tolerance for frequent failures, unlike the zero-tolerance for breakage common in B2B SaaS.

- The inherent unreliability of current AI models compared to the high performance guarantees of SaaS means that for critical business workflows, the "AI killing SaaS" narrative is an overstatement.

- AI should enhance existing SaaS products by making them easier to use, rather than attempting to fully automate or replace core business workflows.

- Successful AI integration in B2B requires products that allow for critical user editing and collaboration, as full end-to-end automation is often an overreach.

- AI features, like "visibility," are commoditizing rapidly, allowing companies to develop and offer them quickly.

- Incumbent companies with large existing revenue bases can offer these commoditized AI features for free, using them for lead generation rather than direct revenue.

- Startups focused solely on easily commoditized AI features face a significant disadvantage against incumbents who can provide these features at no cost due to their existing business models. Sustainable businesses need to build value "downstream of AI visibility."

- AI startups should focus on addressing specific problems for niche buyers instead of developing generalized agents to build more successful businesses.

- Addressing enterprise security and compliance concerns is a major untapped market opportunity that can significantly accelerate AI adoption.

## 00:00 - 04:00 Amplitude's initial skepticism and external pressure regarding AI development

Spenser Skates, CEO of Amplitude, shared the challenges his company faced in adopting AI, particularly as an established tech firm. He noted that larger companies struggle to reorient and integrate new technologies like AI effectively, contrasting this with the agility of newer startups built with AI in mind.

Amplitude itself was initially skeptical about AI's immediate impact, with this sentiment lasting through 2022 and 2023. It wasn't until late 2024 that the company began to seriously consider AI's potential to reshape analytics, prompted by growing external interest.

Skates recounted the frustration of being asked about Amplitude's "AI strategy" by non-technical stakeholders, including salespeople and executives, who didn't fully grasp the technical limitations or the practical implementation challenges of early AI models. This pressure came despite the perceived "terrible" capabilities of AI at the time for their specific use cases.

This led to significant skepticism among Skates, his co-founders, and other team members about the actual transformative effects AI would have on their product and the broader industry. The challenge was in bridging the gap between external hype and internal technical realities.

> Salespeople were like, "Hey, guys, shouldn't you look at this AI thing? Like, isn't that getting hot? Shouldn't you guys do it?" And it's like, "What's your AI strategy?" Yeah, actually, that literally was a question, like, from one of our execs to me, is like, "We gotta get our AI strategy. What's our AI strategy, Spencer?"

## 04:00 - 06:00 Amplitude Kicks Off AI Transformation with Key Hires and Acquisitions

Amplitude initiated its serious push into AI around October 2024, driven by the clear evidence of AI's productivity benefits in software engineering through tools like Cursor, Cloud Code, and Codex. This marked a pivotal moment for the company's strategic direction.

Two major actions underpinned this transformation: the hiring of Wade Chambers, a seasoned engineering leader known for his AI background, and the acquisition of Command AI, a YC company. Both entities were instrumental in bringing the necessary change and expertise to Amplitude.

Wade Chambers brought prior experience in AI from his previous company, along with connections to experts in leveraging model capabilities. Command AI had been developing products to intelligently trigger user guides and chatbots, similar to Intercom Fin, to assist users with support questions and confusion.

Building on these developments, Amplitude has already launched several AI products, including AI Feedback and AI Visibility. The company plans a much larger AI product rollout in December, January, and February, aiming to deliver a

## 06:00 - 10:00 Amplitude's CEO Pivots Organization Towards AI with Dedicated Training and Hackathon

Amplitude, an 800-person company with a 200-person product, engineering, and design organization, faced challenges integrating AI. Despite external pressure and a strong focus on existing product development like experimentation, session replay, and activation, it took about a full year to get the team fully on board and building with AI.

CEO Spenser Skates had a long-standing vision for a self-improving product that dynamically responds to user feedback. Initially, he believed this vision was a decade away, but advancements in AI, particularly on the coding side, made him realize it was much closer. This realization prompted an aggressive push to adopt AI across the company.

To train and galvanize the organization, Amplitude implemented an "AI Week" in June. This event involved existing leaders using the new technology and seeing its possibilities. The week included two days of training for the product, engineering, and design teams, highlighted by a product leader live-coding a dark mode for Amplitude in front of the entire organization, demonstrating AI's immediate capabilities and showing what was possible.

> we've always had this vision at Amplitude of a self-improving product, where you have a product that dynamically responds to your user feedback, so it knows what features you like, when you're getting frustrated and stuck, it knows how to change things based on your input as a user.

## 10:00 - 12:01 AI Product Development Differs from Traditional Customer-Driven SaaS

Traditional SaaS product building follows a clear loop: companies ask customers what they want and will pay for, then prioritize and build those features. This iterative process, exemplified by companies like Amplitude, relies on direct customer feedback to drive product improvements and competitive advantage.

However, building products with AI requires a different approach. Because AI capabilities are often "jagged" and rapidly evolving, customers cannot always articulate what is possible. Asking them what they want might lead to requests like "give me a faster horse," rather than innovative solutions that leverage new AI technologies. Product developers must instead have a technology-first understanding of AI models and how their capabilities can be integrated into the product.

A surprising pattern observed with AI product adoption is that it often occurs top-down within companies. Unlike many new technologies where engineers are early adopters who push for tool usage from the bottom up, AI frequently sees management driving its implementation, even while companies might resist other unproven technologies.

> Give me a faster horse.

## 12:01 - 14:01 The Sam Altman Effect: How Executive Buy-in Drives Top-Down AI Adoption

Spenser Skates attributes the rapid, top-down adoption of AI to Sam Altman's exceptional salesmanship. Altman's ability to articulate an ambitious vision for OpenAI galvanized investors, executives, and even world leaders, convincing them of AI's critical importance. This created a societal consensus that AI matters, even as the technology's capabilities were still developing.

The early and widespread executive buy-in, driven by Altman's influence, often outpaced the actual readiness of AI capabilities. This disconnect led to significant frustration among engineers, who perceived much of the AI discourse as 'grifting'—a lot of talk without concrete, transformative applications. They observed a gap between aspirational visions and practical, deployable solutions.

It wasn't until recently, perhaps the last year, that AI capabilities matured sufficiently to genuinely transform businesses. This shift necessitated a founder-led transformation within companies like Amplitude. Executives, especially founders, had to deeply engage with and utilize the technology themselves to truly understand its possibilities and drive a successful bottom-up integration of AI solutions.

> Man, I see just trem- tremendous what they feel is grifting in AI, where it's a lot of talkers, not many doers.

## 14:01 - 16:01 Amplitude's AI Week drives bottom-up innovation and new product development

Amplitude fostered significant bottom-up innovation through an internal "AI Week," leading to the creation of unplanned but impactful products. Engineers, driven by their own ideas, spearheaded initiatives that contributed substantially to the company's growth.

One notable example is the AI Visibility tool, developed by engineer Leo Jiang, who was originally considering leaving to start his own company. This product, built for free and given away, resulted in a doubling of new signups for Amplitude, showcasing the power of internal entrepreneurial spirit.

Another upcoming product, "Ask AI," a global chat interface launching in January, also emerged from this bottom-up approach. This tool will allow users to interact with AI to pull charts, perform analysis, and understand data within their own contexts.

While these innovations originated from individual engineers, CEO Spenser Skates and Wade's role was to sculpt and integrate these projects into the organization for long-term success. This approach demonstrates a successful transformation that many larger companies struggle to achieve.

> It's all been very bottoms up, and then it's about, you know, for me and for Wade, to, to sculpt, okay, how do we set these up in the organization for success?

## 16:01 - 18:01 Amplitude underwent two reorganizations to integrate AI-native talent and change its leadership mindset.

Amplitude executed two significant reorganizations within a single year to pivot its focus towards AI. This involved transitioning out SaaS-native leaders who were not aligned with the bleeding edge of AI development and replacing them with individuals better suited for the company's future direction.

The transformation included acquiring several Y Combinator companies, such as bringing in founders from Who's, Eric and Frank from Inari, and Enzo and Furci from June. These AI-native founders were then integrated with long-standing Amplitude employees, a combination described as "very special" for fostering innovation.

The fundamental distinction between SaaS-native and AI-native approaches is not an age difference, but rather a difference in mentality. SaaS-native development typically follows an iterative loop: engaging with customers, prioritizing their requests, building features, delivering them, and repeating the process.

In contrast, AI-native development demands taking the current state of the art in a given field and re-imagining how to solve problems through an AI lens. However, many AI-native teams often lack the extensive product and user experience knowledge accumulated over a decade of solving similar problems, leading them to create new interfaces from scratch without leveraging past expertise.

> It's not an age thing per se, it's a mentality thing.

## 18:01 - 20:01 The Mentality Gap: SaaS Reliability Versus AI Iteration

There's a distinct mentality gap between traditional SaaS engineers and those working with AI. SaaS veterans prioritize understanding customer problems, viewing code as a means to an end. They are quick to adapt new technologies if it helps solve a customer's specific issue.

In contrast, interacting with AI-native systems, even internal agents, often involves a high tolerance for failure. Users frequently rephrase questions, switch models (e.g., Gemini, Claude, GPT), or adjust reasoning levels until they get a desired outcome, accepting that the system "fails most of the time still."

This iterative, trial-and-error approach stands in stark contrast to the expectations for B2B SaaS, where even a single failure can lead users to abandon a product. The current state of AI requires users to "rewire their brain" to work with it like "working with a child."

Consequently, claims that AI will entirely replace SaaS are overblown, especially for critical business workflows. Applications like CRMs demand near-perfect guarantees on performance and data persistence, a reliability standard that traditional SaaS excels at delivering and current AI systems often cannot match.

> Man, if it broke even once, like, this thing is piece of crap and I'll never use it again.

## 20:01 - 22:02 AI Enhances Existing SaaS Workflows for Better Usability

Spenser Skates argues against the common misconception that AI will entirely automate business workflows end-to-end. He highlights the critical importance of user editing and intervention in these processes, suggesting that products should be designed to facilitate this collaborative approach rather than aiming for full automation.

Rather than creating entirely new products or replacing existing ones, AI's most effective application in SaaS is to enhance and simplify the use of current offerings. For instance, Amplitude's "Ask AI" chat interface is specifically designed to make their existing analytics product more accessible and user-friendly.

Embracing an "AI-native" approach from the ground up leads to significant benefits. It allows development teams to be faster and more productive, shipping more features. Furthermore, this perspective enables teams to approach problems through an AI-first lens, fostering innovative solutions within existing product lines.

> the goal of, for example, Our, what we're calling Ask AI, which is the chat interface in, in Amplitude, is to make it easier to use the existing product.

## 22:02 - 24:03 Amplitude commits to AI with dedicated teams and self-selection

Amplitude transitioned its AI initiatives from being side projects handled by various individuals to being driven by dedicated teams. This organizational shift aimed to provide focused resources for AI development while ensuring existing teams continued to enhance the core Amplitude product, recognizing the longer cycles of business change.

To foster commitment, Amplitude employed an 'AI Week' and a 'burning the boats' metaphor, signaling an all-in approach to AI transformation. This strategy encouraged employees to naturally self-select into AI-focused roles.

An example of this self-selection is Will Newton, a designer who, despite being spread across multiple projects, chose to deeply focus on a new chat interface. This demonstrates how a clear strategic direction can empower individuals to specialize and contribute significantly to new initiatives.

This approach allowed Amplitude to build focused teams for innovation without disrupting the ongoing development of their main product, balancing future growth with current stability.

## 24:03 - 26:03 Rapid Commoditization of AI Features Challenges Feature-Focused Startups

AI features, such as "visibility" tools, are experiencing rapid commoditization, making them relatively easy for companies to develop and offer. Spenser Skates notes that his company, Amplitude, quickly built and released such a feature.

A significant challenge for feature-focused AI startups is that established incumbents with substantial existing revenue bases, like Amplitude, can afford to offer these commoditized features for free. They leverage these features as powerful lead generation tools rather than a primary revenue driver.

This dynamic undermines the traditional startup advantage, as agility in developing a feature is countered by an incumbent's ability to absorb its cost. For a business to be sustainable, it must be "downstream of AI visibility," building a core business beyond easily replicable AI features. For instance, AirOps focuses on content generation as its main business, using AI visibility as an aspect rather than the whole.

The speaker emphasizes that innovation in this space is moving quickly, but the rapid commoditization means that many visibility businesses will find their offerings available for free from multiple sources, making it difficult to monetize.

> One of the advantages we have is because we have an existing revenue base of hundreds of millions, we can give away this for free.

## 26:03 - 28:03 Google's B2B Products Are Vulnerable to AI Disruption

Spenser Skates identifies Google's B2B products, including email and workspace tools, as significantly vulnerable to AI disruption. He argues that Google is institutionally too slow and conservative to innovate effectively in these areas, creating substantial opportunities for competitors. This slow pace means Google typically only responds to innovation after other companies have already established new paradigms.

He observes this trend already in coding and support, and points to Notion as an example of a competitor successfully challenging Google Docs. Skates believes this pattern makes Google's B2B offerings ripe for disruption, especially in product development and market integration for new solutions.

Skates anticipates a transformative "Cursor moment" in analytics within the next two years. This moment will fundamentally change how people use analytics, making traditional methods seem outdated once AI is integrated. He advises that new AI companies should focus on solving specific problems for particular buyers, rather than attempting to build generalized AI agents, to achieve greater success.

> there's gonna be a cursor moment in analytics in the next two years, no question in my mind, where people are gonna use Analytics with AI and you're gonna be like, "Why did we ever do it the old way?"

## 28:03 - 30:04 Untapped AI Market Opportunities and a 'Uber for Tech Support' Model

Many enterprises struggle to adopt AI, primarily due to significant security and compliance concerns. There's a substantial opportunity for businesses that can directly solve these specific issues, leading to much faster integration of AI products within these organizations.

A practical market gap exists for tech support. The idea is to create a platform, much like Uber, that connects tech-savvy young people who need money with older individuals who are not native to technology but need help setting up and maintaining their devices. This would cater to a clear demand and supply dynamic that currently lacks a scaled solution.

> There has to be a Uber for tech support.

## 30:04 - 32:04 Amplitude's Genesis: Pivoting from Voice Recognition to Product Analytics

Spenser Skates co-founded Sonolight, a voice recognition startup, and participated in YC Winter 2012. Despite an impressive demo day and positive press, the team quickly realized their product's underlying technology was not sufficient for market viability.

During Sonolight's development, the team, being engineers, built their own internal analytics tool. They believed understanding user behavior deeply was crucial for iterating on a better product. This in-house solution proved popular when shared with other companies, who expressed a desire to use it themselves.

This unexpected interest prompted a pivot immediately after YC Demo Day. The team recognized that product analytics presented a "good problem" for their background as algorithms engineers from MIT. Unlike the probabilistic nature of voice recognition, analytics involved building scalable, distributed systems that could deliver precise, verifiable answers efficiently.

Reflecting on this journey, Skates acknowledged the role of luck in finding a problem perfectly suited to their skillset. He also identified a common gap among founders: brilliant engineers often struggled with getting their technology into customers' hands and effectively selling it. This insight motivated him to actively learn and focus on customer acquisition.

> We knew this was a particularly good problem compared to voice recognition. A lot of this AI stuff, frankly, those were probabilistic problems, so you could not get a right answer. Analytics seemed amazing because you build this really scaled distributed system and you can get a right answer, and it's just about doing it better and faster.

## 32:04 - 34:04 Learning B2B Sales as an Engineer

An engineer by background, Spenser admits that even after years in the industry, he still feels like an imposter when it comes to B2B sales. He recognized that for certain complex products, a direct sales approach is indispensable for achieving market adoption and growth, a realization that spurred his willingness to learn.

Spenser's biggest initial misconception about B2B sales was believing it could be learned from books or online resources. He discovered that true proficiency comes only through direct, hands-on experience, combined with the guidance of an expert coach.

To overcome this challenge, he hired Mitch Morando, a seasoned sales executive and coach. Morando would regularly challenge Spenser, pushing him to move beyond simply pitching product features like dashboards and charts, and instead, focus on identifying and addressing the customer's fundamental business pain. This shift in perspective was critical to his development.

> The number one misconception I had was like, 'This would be something you learn out of a book or on a website,' but one, you have to do it, and two, you just want to get someone who's good at coaching you.

## 34:04 - 36:05 Clarifying Your Foundational Purpose Before Launching a Startup

Before embarking on a startup, it is crucial to establish absolute clarity on your objectives and motivations. Many entrepreneurs fail to define their 'why' for building a company, which is a significant oversight.

The speaker emphasizes the importance of understanding your career aspirations and aligning them with a broader mission. He made sure to get clear on what he wanted to achieve professionally and why, before starting his company.

This involves dedicating oneself to a mission that transcends personal gain, aiming to contribute something substantial to humanity. For him, this meant leveraging his skills in building and selling software to achieve this larger purpose.

Once this overarching mission is clear, it provides direction for hyper-focus and allows for the creation of a detailed goal tree, guiding subsequent actions and decisions within the startup.

> You wanna dedicate yourself to a mission that's greater than yourself, and you'd be part of contributing something larger to humanity.

## 36:04 - 38:05 Successful founders defy the rational urge to quit

Starting a company is an extraordinarily emotionally painful journey, and founders will inevitably face moments, even years into their venture, where the desire to quit becomes intense. This feeling is deep-seated and can resurface every few years, challenging one's resolve.

A defining characteristic of successful founders is their ability to push through these periods. There's often a point, typically one to two years in, where from a purely rational standpoint, quitting seems like the logical choice. However, the most successful founders resist this urge and continue on, making their refusal to quit the primary filtering criterion for long-term success.

To navigate these challenging times, founders must have a clear understanding of their fundamental 'why' – the intrinsic motivation that drives them. Relying on external factors like recognition or the promise of financial gain is insufficient to sustain a founder through the inevitable emotional pain and rational arguments to give up. Anchoring back to a deep, personal purpose is crucial for endurance over very long periods.

> there is a point that you get to a year, maybe two years in, where from a rational standpoint, you probably should quit, but for whatever reason, those successful ones don't, and so that is the number one filtering criteria.

## 38:05 - 40:05 Transitioning from a Hands-On Founder to a Large Company CEO

Founders typically lead by example, diving directly into challenging issues like difficult code, product problems, or customer issues, and rallying their teams from the front. This hands-on approach is crucial in the early stages of a company's growth.

However, as a company scales, the role of a leader changes drastically. A CEO can no longer personally lead by example across all departments—sales, marketing, people, product, customers, and press—simply because there are too many demands on their time and attention.

This shift necessitates becoming much more disciplined with time and learning to say no to most requests. The founder might find themselves becoming the kind of big company executive they once criticized for not doing "actual work" and merely judging others' efforts. Embracing this new executive function, despite initial resistance, is essential for continued leadership.

Many successful founder CEOs depart after about a decade precisely because of the profound difficulty in unlearning these ingrained founder habits and adapting to the distinct demands of leading a large organization.

> You become the person you hate. You'd always make fun of big company executives for not doing any work for themselves and just like judging other people's work all the time. But there's a reason for that, and you have to embrace that reason.

## 40:05 - 42:05 Transitioning to a large company executive role requires new skills

The shift from a hands-on founder to a large company executive demands a completely different skill set and toolkit. While founders are often deeply embedded in daily operations, executives at larger firms benefit from established product-market fit and significant resources, such as managing a business with $350 million in revenue and deploying similar amounts annually.

A key aspect of this transition is embracing hierarchy. Unlike the founder's direct involvement, a large company leader must learn to effectively deploy vast resources by leveraging teams and established organizational structures. This allows for a more strategic, less "in the weeds" approach, even if it feels counterintuitive to some founder-mode ideals.

This change is considered the hardest transition by far for many. It involves mastering the art of scaling a business through delegation and strategic resource allocation, a nuanced process often learned through direct experience and coaching rather than abstract management frameworks.

> it's a different, very different toolset and a very different, set of skills in order to, build a business successfully, and that's been the hardest transition by far.

## 42:05 - 44:05 Navigating Late-Stage Company Growth and Public CEO Authenticity

Early-stage company building has established playbooks and guidance, but as companies grow into later stages, clear operational roadmaps become scarce. This creates a challenging environment for founders scaling their organizations without readily available precedents.

For a founder leading a large company, judicious time management is critical. The CEO's time is highly sought after by employees, customers, partners, and investors, requiring deliberate control over one's schedule to maintain focus and effectiveness.

Despite the inherent constraints of being a public company CEO, Spenser is making a conscious effort to be more visible and authentic on public platforms like Twitter. He aims to share his genuine self and convictions, opting against presenting an artificial persona, within the boundaries of his role.

> I don't wanna have to be this other persona that I'm not, or try to represent something I'm not. If I believe and I have conviction in something, I wanna be able to say it.

---

Get podcast briefs for shows you follow: https://podbrew.app/