Podcast Zoo LogoPodcast Zoo
ChatGPT Triggers a New Round of Technological Arms Race

ChatGPT Triggers a New Round of Technological Arms Race

虎嗅·商业有味道
24:33
2023年2月9日
cn
Listen Now

Key

  • ChatGPT: A generative AI chat model developed by OpenAI, which triggered a new wave of the AI boom.
  • AIGC: AI-generated content, which is the basic form of AI directly producing content.
  • Stable Diffusion: An open-source large diffusion-based image generation model that lowers the threshold for AI image generation.
  • Large model: The core technology supporting AIGC, which determines the upper limit of AI capabilities.
  • Business model: A key challenge in the AIGC field, which determines whether a company can develop sustainably in the long term.

Abstract

A technological arms race triggered by ChatGPT is unfolding, and tech giants are flocking to the AIGC field. ChatGPT by OpenAI quickly gained popularity for its powerful conversational abilities, leading people to re - evaluate the upper limit of AI capabilities. Meanwhile, the emergence of open - source models like Stable Diffusion has lowered the usage threshold of AIGC and spawned a number of unicorn companies. However, behind the capital frenzy, the business model of AIGC remains unclear, and many startups are facing profit - making difficulties. Nevertheless, platform - type companies with advantages in technology, data, and computing power, as well as startups focusing on differentiation and complementarity, still have the potential to find their own niches in the AIGC field.

Insights

The podcast content delves deeply into the industry status quo and future trends behind the sensational success of ChatGPT. The rise of AIGC is not just a technological revolution but also an exploration of business models. Although the business model of AIGC is still unclear at present, its application potential in areas such as content creation, marketing, and gaming is huge. In addition, domestic companies still face challenges in large - model research and development and need to increase investment to avoid falling behind in the AI era.

Views

01 "The core of AIGC lies in large models"

AIGC is just an application form of AI, and the large model behind it is the core competitiveness. Companies with powerful large models can take the lead in the AIGC field.

02 "Business model is the key to AIGC development"

Currently, there is a lack of clear and long - term monetization channels in the AIGC field, and many companies are facing profit - making difficulties. Exploring a sustainable business model is the key to AIGC development.

03 "Data and computing power are the cornerstones of AIGC"

The development of AIGC requires a vast amount of data for training and powerful computing power for support. Platform - type companies with advantages in data and computing power are more likely to succeed in the AIGC field.

In - depth Analysis

ChatGPT Triggers an AI Arms Race: Opportunities and Challenges in the AIGC Wave

In early 2023, an AI chat model named ChatGPT quickly became popular, attracting wide attention from the global tech community. This generative AI product developed by OpenAI, with its powerful language understanding and generation capabilities, not only showed people a new height of AI technology but also ignited a new round of AI arms race.

The Emergence of ChatGPT

The appearance of ChatGPT undoubtedly injected a shot in the arm into the long - dormant AI industry. After the AI boom triggered by AlphaGo defeating Lee Sedol in 2016, the AI industry gradually returned to rationality, and many AI algorithm companies were struggling to explore the paths of hardwareization and dataization. The emergence of ChatGPT was like lighting a beacon, bringing the AI industry back into the public eye in the spotlight again.

Different from the previous wave of the AI boom, this time the protagonist has changed from Google's AlphaGo to OpenAI supported by Microsoft. Microsoft CEO Satya Nadella publicly stated that he would integrate OpenAI products into Microsoft's ecosystem and add ChatGPT to the Microsoft suite. Subsequently, Microsoft announced an additional investment of billions of dollars in OpenAI. Meanwhile, OpenAI also announced the latest subscription - based business model for ChatGPT.

Facing the strong rise of ChatGPT, tech giants couldn't sit still. Google urgently launched the conversational AI system Bard, and Baidu officially announced its ongoing project, ERNIE Bot. For a moment, many tech giants entered the fray to seize the high - ground of AIGC.

The Year of AIGC: Coexistence of Opportunities and Challenges

AIGC, or AI - generated content, is the basic form of AI directly producing content. In fact, generative AI is not a new concept, but 2022 was regarded by many AI industry insiders as the Year of AIGC. The popularity of ChatGPT and the emergence of open - source models like Stable Diffusion greatly promoted the development of AIGC.

Stable Diffusion is a large diffusion - based image generation model jointly developed by AI company Stability AI and Runaway. Compared with previous generative adversarial networks, diffusion - based models generate images faster and with better quality. The open - sourcing of Stable Diffusion made AI image - generation capabilities accessible to the general public and became one of the conditions for the phenomenal popularity of AIGC.

On the AIGC track, a number of unicorn companies have emerged, such as OpenAI, Jasper, and Stability AI. These companies have won the favor of capital with their advantages in large models, text - to - text, and text - to - image generation. According to data from research institutions, overseas investors invested at least $1.37 billion in the AIGC track in 2022, completing 78 transactions.

However, behind the capital's enthusiastic pursuit, the business model of AIGC remains unclear. Many startups are facing profit - making difficulties, and some companies have even started to shut down their businesses. For example, the AI image - generation company StalkAI announced that it would stop providing AI image - generation services to users and start business transformation.

Exploration of AIGC Business Models

In the AIGC field, the exploration of business models mainly focuses on the following aspects:

  • API calls: OpenAI attracts application - layer companies like Jasper by opening the APIs of large models such as GPT - 3. These companies optimize the user experience of large models and act as intermediaries between ordinary users and OpenAI. However, this model faces the risk of OpenAI's business adjustment and the challenge of users directly connecting to the underlying models.
  • Subscription services: ChatGPT launched a monthly subscription version at $20, offering value - added services such as no queuing during peak hours, fast response, and priority access to new features and improvements. This model provides ChatGPT with the fastest and most direct way to monetize.
  • B - to - B services: Some AI image - generation companies are starting to provide AI image - generation services to B - end customers such as game studios and art studios. AI image - generation can help customers save a large amount of art costs, but its high randomness limits its applicable scope.
  • C - to - C applications: Companies like Lensar AI and Midjourney attract a large number of C - end users by providing AI image - generation applications. However, it is easy to attract C - end users, but difficult to maintain their freshness.

Challenges Faced by Domestic AIGC Development

Compared with overseas, the development of domestic AIGC still faces some challenges:

  • Lack of large models comparable to GPT: Currently, there is no large model in China comparable to GPT, and OpenAI's API business is not user - friendly to Chinese users.
  • Unclear business model: The business model in the domestic AIGC field is still unclear, and many companies are facing profit - making difficulties.
  • Insufficient data and computing power: The development of AIGC requires a vast amount of data for training and powerful computing power for support. Domestic companies still lack in terms of data and computing power.

Future Outlook

Despite facing many challenges, the development prospects of AIGC are still broad. With the continuous advancement of technology and the improvement of business models, AIGC will play an increasingly important role in areas such as content creation, marketing, and gaming.

For participants in the AIGC track, the future development paths may be divided into three types:

  • Underlying model R & D companies: Focus on the R & D of large models and provide the infrastructure for AI capabilities.
  • Developers of packaged applications: Develop various applications based on large models to meet different user needs.
  • Suppliers providing professional AIGC services to users through applications: Use AIGC applications to provide professional services to users.

In the wave of AIGC, platform companies with data have ecological advantages, while startups focus on differentiation and complementarity. All companies need to make calm judgments in the heat, see clearly the way forward, find their own tracks, and then keep running.

It is worth noting that the development of AIGC also brings some potential risks, such as copyright issues and ethical issues. We need to strengthen the prevention and management of these risks while developing AIGC.

Recommended Episodes

0:000:00