By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
YC Root AccessHow Model ML Is Building the AI Operating System For Finance
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
Great conversations run long. Their best ideas don't have to.
uListen is free on iPhone, iPad, Android, Chrome, and Claude.
See howCHAPTERS
Jump to publisher timestampsChapter summaries
Early intensity in the AI era and what Model ML is building
Chaz describes how the pace at Model ML feels uniquely intense compared to his prior companies, including a first year of extreme hours. He frames Model ML as an “AI operating system for financial services” and argues recent capability shifts made this moment especially decisive.
- •Chaz says the first year was spent working nine-to-nine, seven days a week
- •He positions Model ML as building an AI operating system for financial services
- •He says key agent capabilities changed recently enough that much earlier work felt less useful
- •He notes large-bank procurement cycles are long, adding urgency to timing
From stealth to hypergrowth: team size, capital raised, and global footprint
The conversation moves from intensity to specifics on the company’s current scale. Chaz outlines rapid headcount growth, fundraising totals, and the company’s multi-office presence across financial hubs.
- •Chaz says the business emerged from stealth last February and is about eighteen months old
- •He says headcount grew from roughly five-to-ten people to about two hundred
- •He says they’ve raised over $150M and have eight offices
- •He says hiring is running at roughly twenty people per week
Starting Fat Llama: motivation, background, and the pull toward startups
Gustaf asks why Chaz started Fat Llama, and Chaz describes building companies with his brother Arnie and wanting to create consumer software. He also shares his path from an accounting/finance background into self-taught engineering and the appeal of creativity and “limitless” thinking.
- •Chaz says he built all companies with his younger brother Arnie
- •He describes wanting to build a consumer software business and leaving corporate work
- •He notes he studied accounting and finance and became a self-taught engineer
- •He describes being drawn to a creative, Bay Area-style “limitless” mentality
Getting inspired by YC and the moment they got accepted
Chaz recounts consuming the YC podcast and idolizing YC partners before applying. He describes flying out for interviews and the emotional moment of learning they got in.
- •Chaz says he listened obsessively to the YC podcast
- •He describes meeting YC partners as “meeting celebrities”
- •He recounts flying out for the interview and assuming they didn’t get in
- •He describes getting the acceptance call and being overwhelmed
Social skepticism, staying focused, and what really motivates him
Chaz describes friends being confused or dismissive about entrepreneurship, especially in London at the time. He explains why he keeps pushing anyway—emphasizing competition, learning, and resilience over money.
- •Chaz recalls a friend mocking him for “full-time entrepreneurship” and says friends were largely confused
- •He says they aren’t financially motivated now and would even give away equity
- •He attributes drive to a hyper-competitive family upbringing and love of adrenaline
- •He says startup environments create accelerated learning through freedom to fail
The first month at YC and why founder bonds last
Chaz reflects on building Fat Llama during YC and how being surrounded by founders raises intensity. He explains why batch relationships become lifelong and how high-performing startups differ from “lifestyle startups.”
- •Chaz describes Fat Llama as a consumer marketplace for renting items nearby
- •He says returning to YC repeatedly is valuable because it levels you up around like-minded builders
- •He says batchmates become best friends who “protect each other” over the long term
- •He distinguishes startups trying to win from “lifestyle startups,” and says winning requires moving faster than everyone else
What he carried forward: resilience and scaling communication as headcount explodes
Chaz shares lessons from prior companies: emotional resilience and the need for clear communication structures while scaling rapidly. He highlights how fast growth can leave even “veterans” with only weeks of context.
- •Chaz says startups build emotional resilience and reduce panic in crises
- •He warns rapid scaling demands very clear communication lines
- •He notes that in a jump from ten to two hundred people, “veterans” may have only four weeks’ tenure
- •He emphasizes defining where to ask questions and where knowledge should live
How Model ML began: building for themselves and spotting the due diligence pain
Chaz explains Model ML started as internal tooling while he and Arnie invested their own capital after an exit. Going through transactions made the pain of diligence workflows obvious, and the idea they’d been using themselves became the company.
- •Chaz says after selling their second company they pooled capital and ran a small family-office setup
- •He says early work wasn’t intended for production—it was built for themselves
- •He describes repeatedly pitching ideas to Arnie until Arnie chose the one they were already using
- •He says transaction diligence made the pain clear, especially typing answers to large question banks
What the product does: agents as “digital twins” and an end-to-end banking workflow example
Chaz describes the shift from point solutions to workflow-oriented tools and then to systems resembling a “digital twin” of the user. He gives an example where an agent monitors thousands of companies, triggers analysis, runs modeling, builds a pitch deck, and suggests who to pitch.
- •Chaz says earlier tools were point solutions, then became workflow-oriented, and now function more like digital twins
- •He says a large share of users use the product daily and that use cases are broad
- •He describes automating monitoring, triggering signals, running LBO-style modeling, and generating a pitch deck
- •He says the agent can also recommend who the MD should pitch to, replacing work that used to take a sizable team
How it changes finance work: removing “ridiculous tasks,” shifting juniors to higher-value work
Gustaf asks whether customers do more deals or need fewer people, and Chaz focuses on work quality. He describes eliminating late-night busywork and reallocating junior effort toward more strategic tasks, while noting AI use currently maps to making more money in his category.
- •Chaz describes finance juniors working until 2–3 AM on tasks like ordering logos in slides
- •He argues many repetitive tasks shouldn’t require humans after enough repetition
- •He says juniors can spend more time on value-add strategic work
- •He says in their world, more AI use currently correlates with making more money, while acknowledging the industry may later restructure
Why February mattered: Excel/numerical capability, model progress, and open vs closed source dynamics
Chaz emphasizes a recent inflection where agent systems could finally work properly with numbers and Excel. He also contrasts closed-source model performance concentration with today’s more even landscape and rising open-source capability enabling new cost/performance tradeoffs.
- •Chaz says agent harnesses couldn’t properly work with numbers in Excel until February of this year
- •He encourages “betting on the models improving” as a key dynamic
- •He distinguishes open-source vs closed-source models, noting running on owned GPUs can change what’s economical
- •He says performance is now more evenly split across major closed-source providers and that open-source models are becoming highly capable
Why banks won’t bet on one model: the case for model-agnostic agent harnesses
Gustaf frames the risk of committing to a single model; Chaz agrees and explains why enterprises need the “best model” continuously. He defines “best” as a combination of accuracy, latency, and cost, and argues this requires a model-agnostic layer.
- •Chaz says organizations must keep the best model in users’ hands at all times to stay competitive
- •He defines “best” as multi-dimensional: accuracy, latency, and cost
- •He argues AI should be consumed through model-agnostic agent harnesses
- •He says large financial institutions have massive tech budgets, making the opportunity large
Fundraising realities: what investors bet on, handling rejection, and extracting real feedback
Chaz discusses how early fundraising is largely a bet on founders—work ethic and ambition—more than the product alone. He normalizes frequent rejection, suggests expecting very high “no” rates, and describes pushing investors to give honest feedback quickly to adapt faster.
- •Chaz says early fundraising is mainly a bet on the people, especially willingness to work extremely hard
- •He stresses unusually high ambition as central to raising money
- •He says rejection is normal and can feel like existential failure if you’re not prepared
- •He suggests rejection rates can be extremely high and describes pressing investors for the real reasons to learn faster
Why this likely won’t be built inside banks: iteration speed, culture, and fear-driven urgency
Gustaf asks why banks don’t build these systems internally; Chaz points to faster iteration from being allowed to fail and cultural differences. He argues you can’t easily replicate startup intensity inside a nine-to-five organization, and that existential urgency drives speed and creativity.
- •Chaz says startups can learn by failing and iterate faster than corporates
- •He argues culture is hard to replicate and is often what acquirers want when buying AI startups
- •He says you can’t easily convert a nine-to-five culture into nine-to-nine, seven-day intensity
- •He describes fear of the startup dying as a powerful motivator that improves speed and creativity
Why failed founders get hired and advice for London founders: ambition and starting now
They discuss why founders who tried and failed are attractive hires—because they bring urgency and learned capability. Chaz then gives advice aimed at London/UK founders: increase ambition, don’t overthink, and take advantage of the current opportunity window while young.
- •Chaz says working at startups or scale-ups accelerates learning dramatically and can be reversible career-wise
- •He argues even if it doesn’t work out, people become more capable and employable afterward
- •He advises London founders to be more ambitious and believe they “have the right” to go big
- •He says the current environment offers unusually many opportunities and encourages starting rather than over-optimizing plans
A five-year view and audience Q&A: security, hiring for culture, recruiting, pivoting, and coping with pressure
Chaz forecasts a world where humans do less execution and more judgment and relationship work, supported by multiple proactive agents across communication tools. In Q&A, he addresses security and traceability, hiring for culture, convincing recruits, the balance between perseverance and pivots, and handling pressure and competition (with Gustaf warning about social media).
- •Chaz says future work will emphasize human judgment, relationships, and the value of internal data, with many agents doing execution
- •He describes agents proactively handling tasks across email, Slack, calls, and generating to-do lists
- •On safety, he says deployments can run inside customer infrastructure and that agent traces improve auditability and control
- •On hiring, he recommends optimizing for culture early and building teams that feel like close friends; on recruiting, he frames the choice as building the future vs watching it; on perseverance, he advises persevering when you truly believe customers want it; on pressure, he and Gustaf recommend limiting social media and reframing competition as a positive sign of market demand
Publisher timestamps
- 0:00Nine to nine, seven days a week
- 2:09Starting Fat Lama
- 4:43"You're still doing full-time entrepreneurship?"
- 6:55The first month at YC
- 8:38Why founders become lifelong friends
- 9:56What he took from Fat Lama
- 11:29The idea they'd been building for themselves
- 14:05What the product actually does
- 16:25How it changes the work
- 17:45Why February changed everything
- 19:31Why banks won't bet on one model
- 20:29What investors are actually betting on
- 22:38"19 out of 20 will say no"
- 24:44Why this couldn't come from inside a bank
- 27:42Why failed founders get hired
- 29:30Advice for London founders
- 31:03What the world looks like in five years
- 33:50Q&A: hiring for culture
- 37:03Q&A: convincing people to join
- 40:05Q&A: when to persevere and when to pivot
- 41:52Q&A: coping with the pressure
