Ex Meta Executive Reveals Why Your AI Bill Is About To Explode | Ep. 408 with Cylton Collymore Founder of Sirsi
6/15/202641 min
Daniel and Cylton Collymore dive into one of the biggest questions facing founders right now: what happens when AI becomes essential, but the cost and control of that AI sits in someone else’s cloud. Cylton explains why he believes LLMs are still “book smart, not street smart,” why giving AI agents full access to your computer is like handing your bank account to someone on a second date, and why the future may shift toward local or on-prem AI systems. The conversation also turns personal as Cylton shares how being laid off from Meta after his 50th birthday forced him into Plan A, why Sirsi is his contribution, and why access—not talent—is the real bottleneck holding back the next generation of builders.
Key Discussion Points
Cylton pushes back on the AGI hype, arguing that LLMs present well but are still limited, comparing them to someone who can pass every quiz but struggles with real-world “street smart” reasoning.
He warns founders about giving AI agents full access to their computers, phones, finances, and private data, comparing it to giving a new relationship access to your bank account on the second date.
Cylton explains that agentic systems should support human intelligence, not replace it, especially in roles like cybersecurity where human judgment is needed as the “tiebreaker.”
He shares the deeper vision behind Sirsi: giving people and companies AI efficiency without requiring them to become AI experts or build an AI company from scratch.
The conversation explores why talent is everywhere, from government to Meta to Howard University, and why the future depends on expanding access, encouragement, tools, and visibility for more builders.
Cylton opens up about being laid off from Meta in 2023 and realizing that Plan B was not enough, saying that if you have a Plan B, you do not really have a plan.
He gives founders a blunt lesson on equity: do not give away shares casually, because equity is not candy, it is a claim on your time, effort, and future money.
Cylton explains why he is betting against the purely cloud-centric AI model, arguing that inference costs, token economics, outages, and lack of control will push more AI back to local devices and on-prem systems.
He compares token economics to Chuck E. Cheese at trillion-dollar scale, warning that today’s AI usage is subsidized by investor capital and that prices are already shifting upward.
The episode closes with Cylton pointing toward a hybrid AI future where frontier cloud models still exist, but daily business AI increasingly runs closer to the user through local hardware and private systems.
Takeaways
AI agents are powerful, but founders should not confuse convenience with safety, especially when private data, financial systems, and company operations are involved.
The next AI advantage may not be who has the best prompt, but who controls their infrastructure, costs, data, and inference layer.
Equity is one of the most expensive things a founder can give away, because it represents a claim on years of work, not just a future payout.
Access is the real unlock: more people building with better tools could create a much larger economy and surface talent that traditional systems overlook.
The current AI land grab is being subsidized, and founders should prepare for a world where token costs, cloud dependence, and AI usage become real operating expenses.
Closing Thoughts
Cylton Collymore brings a rare mix of government, cybersecurity, big tech, and founder perspective to the AI conversation. This episode is not anti-AI, it is a warning against using AI blindly before understanding the cost, control, and security tradeoffs. His message to founders is clear: use AI, but do not outsource your company’s future to systems you do not control.
Get up to $200 off Square hardware when you sign up at square.com/go/founder! #squarepod ○ Please make sure to hyperlink this in the YT/Social/Podcast descriptions
Limited Time Offer – Get Huel today with my exclusive offer of 15% OFF online with my code FOUNDER at huel.com/founder. New Customers Only. Thank you to Huel for partnering and supporting our show!
Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
Clips
Transcript preview
First 90 secondsCylton Collymore· Guest0:00
A black guy who's not a software engineer is not gonna get another job in Silicon Valley. It's just the truth. It took me three years of not having a full-time job to get to the point of being able to release Cersei later this year. If you have a plan B, you don't have a plan A.
Daniel Robbins· Host0:17
I've talked to a lot of people, and I don't always learn a lot of things from everyone. I learned a lot today, man.
Cylton Collymore· Guest0:24
The things that I predicted in my thesis in 2023 all actually happened. If you have the ability to have that type of vision before these things are real, I can't imagine what you would do with some money. I don't think AGI is necessarily achievable through LLMs. They're still fairly dumb.
Daniel Robbins· Host0:46
So, Silton, there's so much talk about AGI. Is it a thing? Is it not? I was just watching the Claude, our Anthropic founder, talking about in the next one to two years, we can be at a place that is what he, I don't want to say quoted, but he would say it's unimaginable where we will be at. And I wonder, what does that mean for humanity if the builders don't even know what it means in one to two years?
Cylton Collymore· Guest1:21
Well, he may or may not be right. Know that they released, well, they, they have a, an internal coded