The AI RevolutionSri Viswanath, Vibhor Khanna, Yijia Liang Opens Fullscreen Opens FullscreenOpens Fullscreenverint.comOpens FullscreenOpens FullscreenOpens FullscreenOpens FullscreenOpens FullscreenOpens FullscreenOpens FullscreenOpens FullscreenOpens Fullscreen Opens FullscreenOpens FullscreenOpens FullscreenOpens Fullscreen Opens FullscreenOpens FullscreenOpens FullscreenOpens FullscreenOpens FullscreenOpens FullscreenOpens Fullscreen DisclaimerAI TodayFirst half of 2023Speed of the AdoptionCoatue's broader AI PortfolioAI Founders across stagesHype & ImprovementCoatue ViewComparison: Some enabling technologies become a public goodComparison: AI advancing much faster than previous wavesAI enthusiasmSerious AI buildersChat GPT usageAI is getting better, fasterAI regulationComparison: is AI hype justified with real utility?AI is delivering game-changing valueGitHub Copilot Open-sourceAI is built in the openResult of Open ResearchOpen CollaborationAI Developer communityGitHub for New AI/ML projectsAI is not always openData is finally a new currencyTech EcosystemThe new AI-centric technologyModelsLast 50 yearsNext 50 yearsIntelligence-as-a-serviceWinning in AITalentScaling AI ModelThrough 20222023: GPT-4 remains King$1B training-costHow AI models are madeData ScarcityData QualitySynthetic dataImproving performanceCloud, DC, SemisTraining compute has increasedAI ComputeThe demand for GPUsFollow the GPUsMore GPUs, more expensive serversAI's Holy GrailNew ecosystem of toolsDeveloper ToolsData curation toolsLLM Ops & Vector DBsProblem: Model evaluation is broken todayApplicationsCopilot phase was short-livedAI adoption acceleratingMassive appsAI gameAI StartupsExample: Character AI creating new virtual companionsExample: AI as an autopilot could transform org structuresCoatue ViewThe future top coding languageOn-device AIScaling AIPrivate datasetsInnovation in multimodalityNew advancements in architectureAppendix
