AI search engine Perplexity officially launched its advertising program in the US market this week, creating a new model for search advertising. This highly anticipated AI company has successfully attracted several advertising giants, including Indeed, Whole Foods, Universal McCann, and PMG. Unlike traditional search advertising, Perplexity has innovated an advertising format that integrates into the AI interaction process. Advertisers can naturally incorporate questions in relevant contexts, such as how to use Indeed to optimize job search strategies.

Recently, there has been a craze for 'dog-keeping' on Douyin and Xiaohongshu, but these 'dogs' are not traditional pets; they are an enticing robot dog. Many families are sharing videos of walking their 'dogs' on social platforms, as if showcasing their 'new favorite.' What does this phenomenon actually signify? In the year 2024, known as the 'year of embodied intelligence,' humanoid robots are finally beginning to showcase their 'intelligence' thanks to continuous technological advancements. When mentioning humanoid robots, we cannot help but refer to some star products, such as F.

As a leader in the AI field, Baidu has made significant achievements in the commercialization of AI technology. The daily call volume of its Wenxin large model exceeds 600 million, increasing more than tenfold in six months, with total revenue reaching 33.9 billion yuan, core revenue at 26.7 billion yuan, and operational profit of 5.6 billion yuan, representing a year-on-year increase of 23%, far exceeding market expectations. Baidu's generative AI and foundational models are changing the world, providing solutions for various industries and creating immense value. The user base of Wenxin Yiyan has surpassed 300 million, processing over 1 trillion tokens of text daily.

After Sora's launch, many self-media operators introduced paid courses related to Sora, priced from 50 yuan to 999 yuan, claiming to teach Sora usage techniques and provide creative materials. However, Sora has not been opened to the public. Li Yizhou and other internet celebrities, who claimed to be Tsinghua PhDs, allegedly dissolved their paid group without refunds. Meta scientist Yang Likun questioned Sora's video generation route, declaring it destined to fail. Industry insiders believe that some courses exploit information asymmetry, taking advantage of people's naivety and collecting 'intelligence tax'.

📈 Trends in the Application of Large Models: The article mentions that large models have evolved from the model end to innovations at the application end. 🧑‍💻 Talent Competition in the Large Model Field: The article points out the particularly pronounced talent shortage, with intense competition between large companies and startups. 💰 Commercialization Paths for Large Models: The article analyzes two commercialization paths: cloud services for developers and industry-specific solutions.

Large models themselves do not possess the ability to create value and must rely on application scenarios to realize opportunities and imagination for commercialization. In the future, there may be killer applications suitable for small and medium-sized enterprises to develop AI applications, with the potential for multiple super apps to emerge. Currently, there have been no AI native applications domestically and internationally, but the commercial potential is enormous. The industry is gradually shifting towards the application layer, which is key to the prosperity of the AI industry.

☁️ OpenAI's further commercialization: The article mentions that OpenAI is building a closed-loop ecosystem for GPT commercialization through initiatives like the GPT Store. 💡 AI companies' counterattack strategies: The article analyzes that AI startups should focus on application innovation rather than reinventing the wheel when facing giants like OpenAI. 📈 Strong performance of the GPT series models: The article indicates that the GPT series models currently rank first in traffic among various AI applications.

The saying that large models exhibit certain capabilities only when data, parameter counts, and computing power are large is widespread in the tech community. Competing on model size is unsustainable both commercially and environmentally. The era of seeking sheer size has passed. Although the two paths for large models diverge significantly, their situations are similar. The more promising commercial direction to B2B is currently hindered by high labor and capital costs behind the customizable narrative, preventing business expansion and slow incremental growth. To provide long-term and sustainable services...

iFLYTEK stated that the coding and mathematical problem-solving capabilities of its Spark 3.0 are leading in the industry. They believe that the two major evolution directions for large models are the expansion of model scale and the development of personalized AI assistants. iFLYTEK acknowledged that the greatest pressure from deep cooperation with Huawei stems from their mindset, which can influence the product iteration pace. They forecast that large model business will yield significant commercial benefits next year, reaching a state where revenue growth covers investment by 2025. iFLYTEK believes that there will be obvious differentiation among leading enterprises in the large model field on the scale of 3.5.

OpenAI has achieved an $86 billion valuation through employee stock buybacks, balancing investor returns and public welfare in its equity structure. The company is making progress in commercialization with revenue growth exceeding expectations. OpenAI may be looking to adjust its equity structure to expedite the commercialization process while carefully selecting new investors, with Y Combinator and others potentially entering the scene.

👍 Commercialization of large models: The article mentions that the Baichuan 53B large model has chosen to enter a closed-source commercialization phase. 👍 Large model product matrix: 53B is the third version of the Baichuan large model and the first large model to be commercially closed-source. 🔥 Large models are a focus: The outside world is paying attention to Baichuan Intelligence's strategic choice between open-source and closed-source.

🔥OpenAI Valuation: OpenAI's valuation has surged nearly 3 times from $29 billion seven months ago to around $80-$90 billion. 💰OpenAI Revenue: OpenAI expects to reach $1 billion in revenue this year, with the potential to generate billions in revenue in 2024. 📈Large Model Commercialization: OpenAI is leveraging its technological advantage to accelerate commercialization.

1. Several domestic large models have officially launched and opened for trial, rekindling enthusiasm in the artificial intelligence market. 2. Tencent and Meituan are investing in self-developed large models while supporting startups. 3. Technological strength and commercialization capabilities are key for large model companies to secure investments.

OpenAI expects to generate over 1 billion dollars in revenue within the next 12 months through the sale of artificial intelligence software and its computing power. OpenAI's business model mainly includes charging per API call and a subscription-based model. The relationship between OpenAI and Microsoft is complex, and after commercialization, both parties may face conflicts of interest.

OpenAI has acquired digital product company Global Illumination, venturing into the gaming sector. OpenAI faces challenges in commercialization and needs to address high training costs. Acquiring a gaming company may help OpenAI alleviate commercialization pressures.

People often overestimate the short-term impact of technology while underestimating its long-term contributions to social development. Huang Xuedong believes that in the commercialization of AI, there is a need to focus on the integration and interoperability capabilities of models. The future trend will center around large models with joint development of multimodal functionalities.