By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
The Twenty Minute VCGoogle Loses $500B in 12 Hours | DoorDash Buys Instacart for $17B | Kalshi Raises $1B
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
CHAPTERS
- 0:07 – 2:11
OpenAI’s down-round sparks ‘AI wave cracking’ debate
Harry lays out OpenAI’s reduced valuation and smaller raise, prompting a discussion about whether this signals broader weakness in AI. Rory argues it’s straightforward relative pricing versus a larger, faster-growing, liquid public alternative, not a narrative shift about AI itself. Jason agrees AI is underappreciated but says it may need to drop further amid volatility.
- •OpenAI’s valuation and round size are described as reduced, with partial commitments
- •Rory rejects the “AI wave cracking” storyline and frames it as normal relative valuation and liquidity discounting
- •Jason says bumps are expected and the valuation may need to come down more
- •The group frames current conditions as volatility rather than a definitive turning point
- 2:11 – 4:58
Supply-side ‘price discovery’ in AI: debt costs, IPO wobble, and market nerves
Jason and Rory describe a market environment where lending terms and private valuations are being recalibrated. They cite expensive debt and underwriting signals as evidence of wobbliness, especially heading into IPOs. Rory distinguishes between supply-side repricing and whether demand ultimately holds up.
- •Jason points to large market swings and heightened volatility as reasons for wobbliness
- •Debt markets are described as signaling risk, including high yields and guarantees
- •Rory frames this as price/risk discovery on the supply side of the AI trade
- •They argue the key determinant is whether AI demand remains strong or weak
- 4:58 – 6:45
Alphabet’s ~$500B drop: search angst, capex concerns, and narrative whiplash
The panel reacts to Alphabet’s sharp decline and debates what investors are repricing. Rory attributes it to search cannibalization fears alongside broader market and capex angst, despite strong cloud growth. Jason describes a market where even small misses trigger outsized selloffs because investors are “nervous.”
- •Harry highlights Alphabet’s stock drop alongside cloud growth and slower search growth
- •Rory cites search cannibalization worries plus capex and general market angst
- •Rory notes fast-shifting sentiment swings around big tech leadership narratives
- •Jason argues markets are hypersensitive and punishing even slight shortfalls
- 6:45 – 8:03
Anthropic IPO pressure: decelerating bookings and intense model pricing competition
Jason predicts Anthropic could face pricing pressure and investor nerves about decelerating quarter-over-quarter momentum. He cites cheaper competing offerings and falling effective prices as the core issue rather than demand collapse. The discussion emphasizes how difficult it is to keep growth blazing amid rapid deflation in model costs.
- •Jason expects IPO pricing pressure driven by growth deceleration concerns
- •He argues pricing pressure is the simplest explanation for softer incremental momentum
- •Competitive discounting and cheaper “latest models” are framed as forcing deflation
- •They debate how much “dominance” is required to hold a premium in a crowded market
- 8:03 – 13:58
Interpreting the numbers: absolute vs percentage growth and what investors are betting on
Rory walks through how the same revenue trajectory can look like either extraordinary growth or sharp deceleration depending on the lens. Jason stresses that even small slowdowns at the margin can change investment decisions. They conclude the key question is whether deceleration continues or stabilizes at still-remarkable rates.
- •Rory explains how equal absolute additions can imply lower percentage growth as the base grows
- •He shows how the same facts support both bullish and bearish interpretations
- •Jason argues that slower incremental additions would deter him in a startup context
- •They frame the “only discussion” as the forward bet on the next growth step-down
- 13:58 – 16:11
AI model deflation in practice: a personal cost drop and ‘almost impossible’ forecasting
Jason shares an anecdote from building his own product where switching models reduced costs dramatically while improving quality. The group uses it to illustrate why forecasting AI economics is unusually difficult. Harry links this to broader venture and public market volatility driven by uncertainty about the medium-term path.
- •Jason reports a major effective cost decrease for the same tasks after switching models
- •He emphasizes quality improved while cost fell, changing his view of real-world deflation
- •They argue rapid capability and pricing shifts make 12–60 month predictions hard
- •Harry ties unpredictability to market volatility and binary venture outcomes
- 16:11 – 19:48
Canva and ‘quiet obsolescence’: how agents and AI tools reshape legacy software demand
Jason says he stopped using Canva after many years and suggests some products may have missed their window. Rory argues software that stops moving forward can drift into terminal value over time. Jason adds he’s surprised how broadly some “core businesses” seem impacted even outside power-user circles.
- •Jason describes churning Canva and speculates about its trajectory
- •Rory says software that isn’t moving forward tends toward decline and terminal valuation
- •Jason expresses surprise that mainstream demand appears to be weakening across some products
- •They discuss how strong AI image tools and agents can undercut established workflows
- 19:48 – 25:36
Consumer agents and unit economics: persistent compute costs, subsidies, and power users
The conversation turns to whether consumer agent products can sustain their economics, especially with persistent VMs. Jason questions whether pricing can work if heavy users cost far more than subscription revenue. Rory argues many digital products begin with imperfect economics, relying on cost curves to improve before capital runs out.
- •Jason questions whether persistent-agent compute can be economical at typical subscription prices
- •Harry argues light users can subsidize heavy users via blended economics
- •Rory says products often scale before unit economics are perfect, betting on declining costs
- •They debate whether subsidies end before cost reductions arrive
- 25:36 – 27:52
DoorDash buys Instacart for $17B: consolidation logic and integration surprises
Harry introduces DoorDash’s acquisition of Instacart and frames it as a major deal that would normally dominate the news cycle. Rory says it makes strategic sense and may be a ‘do it now’ move. Jason notes he’s been surprised that some delivery-network acquisitions can be accretive rather than destructive.
- •Harry outlines the deal size and positions it as a second-place consolidation outcome
- •Rory says the transaction broadly makes sense and timing may matter
- •Jason cites past surprises where delivery-network acquisitions proved accretive
- •They discuss how scale and market share can strengthen network dominance if approvals clear
- 27:52 – 31:42
Are agents an existential risk to delivery platforms? Disintermediation and ‘competitor #3 is nobody’
Harry cites DoorDash’s claim that agent-driven orders are currently tiny relative to total volume, and the panel debates what that means. Jason argues the comment is self-serving and that agents could meaningfully disintermediate platforms by routing around fees. He frames DoorDash’s “number three competitor” as restaurants’ direct options enabled by agents.
- •Harry questions whether low agent-order counts imply limited disruption or just early days
- •Jason calls the disclosure self-serving and says all app platforms face disruption risk
- •They discuss agents optimizing across platforms and direct restaurant fulfillment to avoid fees
- •Harry argues discovery and browsing are core behaviors, limiting full automation for many users
- 31:42 – 36:19
Kalshi vs Cohere: momentum financing, regulation battles, and ‘sovereign AI’ subsidy dynamics
Rory calls Kalshi a ‘money deal’ driven by rapid growth despite regulatory conflict, while Cohere’s round is framed as government-backed strategic positioning. The group contrasts a business pushing through regulation with one buoyed by national interest even if economics are uncertain. They debate whether governments can fund outcomes without private-sector pull-through.
- •Rory describes Kalshi as working and benefiting from momentum despite regulatory pressure
- •Cohere is framed as a sovereign AI play supported by government funding motives
- •They argue governments can finance for a time but can’t force broad private adoption
- •Jason critiques non-disclosed revenue as a negative signal in the current environment
- 36:19 – 38:36
Risk, regulation, and crypto parallels: Kalshi, Binance, and ‘game on’ risk-taking
Jason draws a parallel between investors backing companies under legal scrutiny and how outcomes can hinge on timing. The group discusses Binance’s ability to cover a major loss as evidence of strong economics, while also noting uncertainty about transparency. The takeaway is that venture risk tolerance can look inconsistent yet be rewarded when momentum holds.
- •Jason wonders how different timing could have changed outcomes for similar-risk companies
- •They discuss Binance covering a large loss as signaling a highly profitable business
- •Jason argues transparency and custodial ethics separate outcomes in crypto blowups
- •They reflect on how investors continue to fund what appears to be working despite legal risk
- 38:36 – 41:23
AI pen-testing tools as cyber weapons: ‘defense’ code turned into scalable attack
Jason describes a new open-source pen-testing tool that can be repurposed to attack systems using LLMs at scale. Rory agrees the dual-use dynamic is central and warns cyber events could become startlingly disruptive. Harry adds examples of real-world infrastructure vulnerability and notes how quickly stakes can escalate beyond isolated hacks.
- •Jason explains how a pen-testing tool can be flipped into an offensive hacking framework
- •They emphasize LLMs enabling scalable targeting and automation of attacks
- •Rory predicts major cyber dynamics could surprise the public within a year
- •Harry discusses potential spillover into critical infrastructure beyond consumer devices
- 41:23 – 52:32
Choosing between Cognition, Baseten, and Serval: valuation, revenue velocity, and TAM expansion risk
Harry asks Jason to pick among three fast-reraising companies, forcing a tradeoff between scale, upside, and market structure. Jason argues coding agents are consistently working and leans toward Cognition despite high valuation, while Harry prefers Serval for multiple potential and product momentum. The group debates whether Serval expands the TAM or just fights incumbents, and whether inference is attractive despite intense competition.
- •Jason favors Cognition on momentum and the proven pull of coding agents despite valuation concerns
- •Harry prefers Serval for perceived multiple upside and strong enterprise adoption
- •They debate TAM expansion versus competing within an incumbent’s market footprint
- •Rory argues inference offers unmatched revenue velocity but faces competition and commoditization risk
- 52:32 – 59:02
Can AI reinvent the law firm? Weil’s collapse, Crosby’s bid, and the limits of ‘human plus AI’ models
The discussion shifts to a major law firm collapsing after key partner departures and an AI-oriented firm bidding for assets. Rory argues elite law is partner-driven, while standardized work is more amenable to automation and brand-agnostic delivery. Jason questions whether a venture-scale outcome is feasible given legal industry compensation dynamics and incentives.
- •They describe how law firms can unravel when key rainmakers and books of business leave
- •Rory argues transactional legal work is more replaceable and could fit an AI-enabled model
- •Jason questions whether lawyers will trade partner economics for equity-driven outcomes
- •They agree a business can exist, but debate whether it can reach very large outcomes
- 59:02 – 1:08:32
Robotics at $20B and why venture needs M&A: long deployment cycles, big-tech TAM narratives, and antitrust worries
Rory contrasts high valuations for low-revenue robotics with low valuations for high-revenue mature software, arguing the market is future-driven. He’s bullish on model capability improvements but notes robotics deployment takes longer and requires far more installed robots to justify huge valuations. Jason closes by emphasizing that venture depends on steady M&A, expressing concern that scrutiny of deals could slow exits even as large acquirers remain active.
- •Rory says revenue multiples matter less than future narratives, citing robotics versus mature SaaS
- •He praises robotics model breakthroughs but warns deployment timelines are longer
- •He suggests big foundation-model companies may buy or build robotics to claim a larger TAM
- •Jason argues venture needs frequent M&A, worries about deal scrutiny, and notes acquirers often want $100M+ revenue before buying
