AI不是短跑比赛

Core Viewpoint - 2023 is recognized as the "year of effectiveness" for AI, with 2024 expected to be the "year of application" [2] - The AI large model industry has not yet entered a true "explosion period," with user scale and app installations increasing but revenue conversion still lagging [2][3] Group 1: AI Development and Market Dynamics - Baidu's AI strategy has transitioned from sowing to nurturing, focusing on small successes rather than immediate harvests [3][4] - Historical evidence suggests that the transition from technology accumulation to explosion cannot be strictly planned, with each harvest being a reward for long-term commitment [7] - The initial hype surrounding ChatGPT is fading, revealing limitations in user engagement and interaction duration, with average user interaction lasting only 6-15 minutes [10] Group 2: Competitive Landscape - OpenAI's challenges highlight the fragility of newly created application scenarios, as seen with Sora 2 facing copyright and harmful content controversies [11] - Gemini's growth has outpaced ChatGPT, with market share increasing from 6.4% to 12.9%, while ChatGPT's share has dropped from 87% to 74% [11][12] - Baidu's AI search has achieved 3.82 billion monthly active users, reflecting an 18.6% quarter-over-quarter growth [12] Group 3: Long-term Value and Strategy - The true value of AI lies in its integration with time, emphasizing the importance of long-term strategies over short-term gains [13][15] - Baidu aims to create a comprehensive AI stack rather than focusing on single-point applications, avoiding the pitfalls of the previous mobile internet era [16][25] - Baidu's autonomous driving service, "萝卜快跑," has accumulated over 240 million kilometers of driving, showcasing significant potential in a validated demand scenario [18] Group 4: Internalization of AI Capabilities - Baidu's concept of "internalizing AI capabilities" emphasizes the need for AI to be integrated as a native ability across industries, driving productivity and economic growth [20] - The AI industry faces challenges with low investment efficiency, as evidenced by the inflated valuations of AI startups and the lack of positive returns for many enterprises adopting AI tools [21][23] - The understanding and effective utilization of AI is a longer process than anticipated, indicating that the "big gathering" of AI is still on the way [24][26]