Deck presents the top 10 predictions for the next 12 months.
A sovereign country will invest over $10 billion in a major US AI lab, triggering national security reviews. A $10B+ investment from a sovereign state into a US large AI lab invokes national security review.
My personal view: I think US AI companies will be very cautious about accepting foreign capital, especially funds from sovereign states. Especially on October 28, the US Department of Treasury issued new executive orders that further tightened regulations, which will also increase the complexity and uncertainty of such transactions. You can refer to this policy release: https://home.treasury.gov/news/press-releases/jy2687.
An app or website entirely created independently by someone with no coding ability will become popular and rank in the top 100 in the App Store. An app or website created solely by someone with no coding ability will go viral (e.g. App Store Top-100).
My personal view: I strongly agree with this point. For example, Leo has developed an English learning website all by himself using a bunch of AI agents, which has become one of the products I use most frequently in my daily life, and it's extremely practical. Here is his website: https://enspiring.ai/.
The frontier laboratory made significant adjustments to its data collection practices as some related cases entered the trial stage. Frontier labs implement meaningful changes to data collection practices after cases begin reaching trial.
My personal opinion: I don't know much about the legal issues regarding data collection.
The early implementation of Europe's 《AI Act》 was less stringent than expected because lawmakers were concerned that the act might be too harsh. Early EU AI Act implementation ends up softer than anticipated after lawmakers worry they’ve overreached.
My personal opinion: Although my understanding of European laws is limited, indeed, if the regulation had been even more rigid, Europe might have struggled to keep up with this rapidly evolving era. Maintaining a moderate degree of flexibility could potentially be more conducive to innovation.
Open-source alternatives outperformed OpenAI o1 in several reasoning benchmark tests. An open source alternative to OpenAI o1 surpasses it across a range of reasoning benchmarks.
In my personal opinion, it is very likely that open-source solutions could surpass o1. As far as I know, the training of GPT-5 encountered some challenges, which led to the temporary release of the o1 version that was originally not planned for public release. If it's only about optimizations like Chain of Thought (CoT) and other reasoning improvements, open-source models may achieve or even surpass o1 in certain aspects. However, there are rumors that OpenAI is preparing more innovations, so even if open-source models surpass o1 in the short term, OpenAI might quickly regain its advantage through other new products or updates.
Competitors have failed to substantially impact NVIDIA's market position. Challengers fail to make any meaningful dent in NVIDIA’s market position.
In my personal opinion: I am not very familiar with the chip market.
The enthusiasm for investing in humanoid robots will gradually decline due to the mismatch between products and market demand. Levels of investment in humanoids will trail off, as companies struggle to achieve product-market fit.
In my personal opinion: Although I'm not very familiar with the humanoid robot market, after watching Elon Musk's presentation and seeing a humanoid robot priced lower than a car, it was indeed very inspiring:
Apple's strong results in on-device AI research will accelerate the momentum of on-device AI development. Strong results from Apple’s on-device research accelerates momentum around personal on-device AI.
My personal view: I have previously watched Apple's launch events and am indeed looking forward to its progress in on-device AI. This has the potential to significantly enhance the intelligence of devices and the experience of personalized applications.
Research papers generated by AI scientists will be accepted at major machine learning conferences or workshops. A research paper generated by an AI Scientist is accepted at a major ML conference or workshop.
In my personal opinion: Although I am not very familiar with the application of AI scientists in the field of machine learning academia, judging from its applications in fields such as biology and proteins, AI indeed performs exceptionally well as a research assistant. As the "Copilot" for scientific research work, it is worth looking forward to AI achieving more breakthroughs in various academic fields.
Electronic games based on interactions with generative AI elements will achieve explosive success. A video game based around interacting with GenAI-based elements will achieve break-out status.
In my personal view: The large-scale outbreak of this type of game is undoubtedly just a matter of time, but whether it will happen within the next 12 months remains to be seen. With the increase in people's leisure time and the rise in demand for relaxation, interactive experiences will become even more important. The controllability of AI continues to improve, and the interactive experience between users and generative AI content will become increasingly rich. It can be foreseen that more such games will emerge and gain widespread popularity in the future. I once saw an interesting experiment: