Ron Josey

Huffman described the shape of the relationships and the value framework but gave no renewal timing, terms or economics.

Huffman answered specifically, naming feed improvements, onboarding and web-to-app conversion, and volunteering that Reddit's feed technology is behind state-of-the-art.

Steve tied bot verification, Passkey-based login and human verification together as overlapping transparency and defense work.

Steve said streamlining onboarding improves retention and Reddit will invest in ML to make the cold-start feed better.

Steve said Q3 growth was a balance of organic and paid with external search flat, and Reddit exited Q3 above its average into Q4.

Jen said growth is led by well-retained existing advertisers, a new global accounts team, third-party integrations and cross-funnel adoption.

Steve said Reddit is unifying traditional search and Reddit Answers into one front-and-center in-app experience for seekers and scrollers.

Steve gave an early April read of high-teens DAU growth and framed Google-driven bumps as expected but not a long-term risk.

Jen said there is a lot of uncertainty but it is mostly business as usual, with Reddit helping advertisers navigate volatility.

Steve described Answers as an early beta testing corpus depth and answer quality, calling the start promising.

Steve said such swings happen regularly, mainly hit logged-out users, and traffic recovered, but did not specify the change.

Chesky described a new AI-driven pricing model as one of the largest growth levers, well above Reserve Now Pay Later, while experiences remain a longer-horizon, small-base opportunity.

Brian explained multiple drivers: aggressive push to download app, increased notification opt-ins, email hooks and App Store ranking improvements.

Brian said it is very early but Airbnb must move at the speed of AI. Leaders need to be more hands-on with less room for pure people managers.

Brian outlined three-phase AI search: Phase 1 natural language input with AI-generated search cards, Phase 2 multi-turn conversational, Phase 3 unified AI concierge integrating search, messaging, and customer service. Paris experiences revealed three audience segments with different supply needs.

Brian described personalized carousels, post-booking entry points, and dedicated Paris team doing rapid iteration. Loyalty program under active consideration but not imminent; would be novel, not points-based.

Ellie said May 13 launch will have modest Q2 revenue impact with expense building over the year. Affordability positioned as natural hedge with ability to merchandise lower-cost nearby listings.

Brian detailed personalized welcome tour, suggested destinations, shorter checkout flow driving conversion. Ellie quantified product optimization lift at approximately 200 basis points.

Glenn described a grind-it-out approach of product, inventory and marketing improvements; Ewout said the incremental $100M savings came mostly from procurement.

Glenn explained multi-brand AI strategy: teams share learnings across brands, best approaches get scaled. Also mentioned undisclosed startup-type AI projects in development.

Glenn confirmed small-sample conversion uplift from Penny. Ewout added metrics beyond conversion: faster search, shorter path to booking, lower cancellation rates, and positive customer satisfaction.

Ewout cited better conversion, lower cancellation rates, reduced customer service contact rates, faster handling times, higher satisfaction, and absolute cost declines despite 10% volume growth.

Glenn acknowledged US brand awareness needs improvement, emphasized continuing investment in product and brand advertising as the path to higher direct share.

Glenn declined to share specific metrics but confirmed natural language search remains live because it's performing positively, and cited broader GenAI wins across brands.

Ewout described 'double discipline', driving operating leverage while reinvesting $170M in GenAI, fintech, and verticals for higher future top-line growth.

Confirmed the new tools are not fully deployed, framed it as remaining opportunity, and cited compounding customer-care cost declines from AI.

Continual improvement in conversion across all funnel definitions. Opportunity remains in understanding and trust. Tests confirm clear elasticity relationships with speed and price.

ADESA sites currently at less than half the utilization of Carvana sites. Catching up is asymptotic and will take time, but on plan or slightly better.

Acknowledged tariff uncertainty. Small demand pull-forward then trough then stabilization. Focus remains on relative competitive positioning, not chasing macro.

Walked through EBITDA margin economics: 11.5% base + fixed cost leverage + older market benefits + fundamental gains well exceeds 13.5% target.

Highlighted Sebastian chatbot as first application. Cited 4M+ transaction data set and vertical integration as unique AI advantages.

Framed excess inventory depreciation as a customer acquisition cost that can be compared to other channels. Team has done incredible job scaling.

Zuckerberg said 2026 is key for establishing agents as the product vehicle, but model improvement will continue for a very long time across short, medium, and long term.

Li pushed back on the diminishing-returns thesis, describing a multi-pronged roadmap of data infrastructure improvements, model complexity scaling, and LLM-based tuning.

Susan Li described GEM as the first recommendation architecture scaling like LLMs, with 2026 plans to double the GEM cluster and expand data/complexity.

Susan Li quantified a 14% lower cost-per-lead for Advantage+ lead campaigns and flagged meaningful adoption headroom across existing advertisers.

Zuckerberg pointed to model upgrades driving engagement but declined to specify the roadmap.

Susan Li gave rich color on use cases (info seeking, emotional support, humor) and flagged memory/personalization as the 2025 roadmap.

Zuckerberg acknowledged possible near-term business trade-offs to refocus Facebook on cultural influence but gave no specifics.

William detailed segment-specific selling and measurement, and taking time to get Pinterest Assistant product-market fit right.

Bill called tvScientific the first move into off-platform supply in fast-growing connected TV, with the transformation already in flight.

Bill framed Pinterest as already delivering the agentic promise proactively; Julia said UCan tariff pullback partially recovered.

Bill cited Adobe data on Gen Z search adoption and the platform's visual-search resonance across geographies.

Julia named financial services, technology, telecom and entertainment as underpenetrated verticals; AI code up 10 points to 25%.

Bill cited largest advertisers at over 80% lower-funnel spend and room to grow share of wallet in the next tranche.