2024 AI Six Major Predictions: OpenAI's Transformation, Apple's Entry... The AI battlefield will see a major reshuffle

In 2023, the AI wave led by OpenAI's OpenAI Large Model swept the business world at an unprecedented speed. After a year, the market has gradually realized the power and limitations of AI, and 2024 will be the year when the hype becomes a reality.

According to the well-known technology media TechCrunch, the following trends are expected to emerge in the AI industry next year:

01 OpenAI will become a product company

After the power struggle and leadership adjustments in November, OpenAI will transform into a company that is more focused on "launching products," and the comprehensive influence of Sam Altman will gradually manifest at all levels of the company.

TechCrunch points out that the OpenAI Store is an example. The OpenAI Store, which integrates various AI tools, is very similar to Apple's App Store and is expected to surpass open-source tools like Hugging Face to become the default place for users to obtain AI tools.

With the cautiousness of the previous board of directors giving way to market pursuit of speed and development, OpenAI may launch more heavyweight products similar to the OpenAI Store next year.

02 Agents, video generation, and music generation moving from niche to mainstream

In 2024, some niche applications of AI models will surpass the "uh" status, including agent-based models and generative multimedia.

If AI is to take on more complex tasks, it may need access to your spreadsheets, ticketing interfaces, transportation applications, etc., which are called "agents." TechCrunch believes that in 2024, agents may not make any major breakthroughs, but agent-based large models will be more advanced than last year, and some tedious processes (such as submitting insurance claims) are expected to be improved with AI.

Video and audio generation will also find their niche markets, and in the hands of skilled creators, the lack of realism in AI-generated works is not a problem. AI-generated videos can have interesting applications. Similarly, generative music models are likely to be applied to large-scale productions such as games, allowing professional musicians to create endless soundtracks using these tools.

03 Hybrid small models may have an advantage

With the increase in computing power, the capabilities of large language models are also improving accordingly. However, in 2024, the technological trend may become clearer: a single overall model may not be as efficient as a stack of multiple models.

TechCrunch points out that the diminishing returns of training parameters are already very clear. Although technically, a large model with 500 billion parameters may indeed have comprehensive capabilities surpassing small models, the training cost is too high and it is too cumbersome for updates and iterations.

In contrast, the same computing resources can be used to train multiple small models, and the combination of these small models may not be inferior to large models in terms of generating results, and it is easier to perform scattered updates.

04 Many companies will give up large-scale AI applications

After a year of AI hype, it will be difficult to continue in 2024. Companies that have invested heavily in AI systems will face market questioning and scrutiny at their earnings meetings, and it is very likely that the results will not meet expectations.

Due to the inability to prove the cost-effectiveness and risk of AI tools, TechCrunch predicts that many companies will give up large-scale AI applications in their business.

The future industry competition will become more fierce. Although the capabilities of AI will continue to grow and progress, not all products from 2023 will survive. Companies that are wavering in the wave will fall and be swallowed up, leading to a wave of consolidation in the AI industry.

05 Apple is about to join the battle

Apple's usual approach is to wait, observe, and learn from the failures of other companies, and then join in with carefully crafted products to lead the technological trends.

TechCrunch believes that next year is a good opportunity for Apple to showcase its capabilities in the field of artificial intelligence. If the waiting time is too long, its competitors may eat into the market. Moreover, AI technology has matured, and the time for Apple to demonstrate its product power has arrived.

It is still difficult to say whether Apple's AI service will be an improved version of Siri or a completely new service. However, considering Apple's vast ecosystem, vast amount of data, and almost unlimited R&D funds, TechCrunch predicts that Apple's AI service will definitely have multimodal understanding capabilities and be able to perform complex proxy functions. For example, "Siri, book a table for four at a sushi restaurant in the city center around 7 o'clock and arrange a car to pick us up."

06 A new AI compliance industry may emerge

Major moves such as the EU's "Artificial Intelligence Act" may change the way the industry operates, but these regulatory rules often take effect slowly, allowing companies to adapt to the new rules overnight.

This also means that, apart from companies that voluntarily make changes, these laws may not show any effect for a long time. Tech companies may continue to procrastinate, and there will also be companies using litigation to challenge these laws.

TechCrunch predicts that this situation will lead to the flourishing development of the AI compliance industry. After all, companies that have already invested heavily in large-scale model development and computing power probably do not want to spend additional money on resolving litigation, which is highly uncertain.