This is good news for OpenAI. For Tesla shareholders, this will be unfortunate news. And for those concerned about AI safety, this will be a disappointing news, as it virtually ensures that the US government will take a hands-off approach to AI regulation during the Trump administration.
04. AI Agents Will Become Mainstream
Imagine a world where you no longer need to directly interact with the internet. Whenever you need to manage subscriptions, pay bills, book doctor appointments, order things on Amazon, make restaurant reservations, or complete any other tedious online tasks, you can simply instruct your AI assistant to do it for you.
The concept of "internet agents" has existed for years. If such a product existed and worked properly, it would undoubtedly be a huge success.
However, there is currently no generally functioning internet agent on the market.
Companies like Adept, even with a blue-chip founding team and hundreds of millions in funding, have yet to realize their vision.
Next year will be the year when internet agents finally start to work well and become mainstream. The continuous progress of language and visual foundation models, combined with recent breakthroughs in "second-system thinking" capabilities due to new reasoning models and reasoning time computation, will mean that internet agents are ready to enter their golden age.
In other words, Adept's idea was right, just premature. In startups, as in life, timing is everything.
Internet agents will find various valuable enterprise use cases, but we believe the biggest near-term market opportunity for internet agents will be in the consumer space.
Despite the recent AI hype, there are relatively few AI-native applications that have become mainstream consumer applications, apart from ChatGPT.
Internet agents will change this, becoming the next true "killer app" in the consumer AI space.
05. The Idea of Placing AI Data Centers in Space Will Be Realized
In 2023, the key physical resource constraining AI development is GPU chips. In 2024, it becomes power and data centers.
In 2024, there will be hardly any story more captivating than the enormous and rapidly growing demand for energy as AI rushes to build more and more AI data centers.
Due to the explosive growth of AI, global data center power demand is expected to double between 2023 and 2026, after remaining flat for decades. In the US, data center electricity consumption is projected to approach 10% of total electricity consumption by 2030, up from just 3% in 2022.
Today's energy systems are simply not equipped to handle the massive surge in demand from AI workloads. A historic collision is about to happen between our energy grid and our computing infrastructure, both worth trillions of dollars.
As a potential solution to this dilemma, nuclear power has seen a resurgence this year. Nuclear power is in many ways an ideal energy source for AI: it is zero-carbon, available 24/7, and effectively limitless.
But realistically, new energy sources cannot solve this problem before the 2030s, due to the long lead times for research, project development, and regulation. This applies to traditional nuclear fission plants, next-generation "small modular reactors" (SMRs), and nuclear fusion power plants.
Next year, a very unconventional new idea to address this challenge will emerge and attract real resources: placing AI data centers in space.
AI data centers in space, at first glance, sounds like a bad joke, a venture capitalist trying to mash too many startup buzzwords together.
But in fact, it may make sense.
The biggest bottleneck to rapidly building more data centers on Earth is acquiring the necessary power. Computational clusters in orbit could enjoy free, limitless, zero-carbon power 24/7: the sun is always shining in space.
Another key advantage of placing computing in space is that it solves the cooling problem.
One of the major engineering hurdles in building more powerful AI data centers is that running many GPUs in a confined space gets extremely hot, and high temperatures can damage or destroy computing equipment.
Data center developers are adopting expensive and unproven methods like liquid immersion cooling to try to solve this. But space is extremely cold, and any heat generated by computing activity would dissipate harmlessly immediately.
Of course, there are many practical challenges to be solved. An obvious one is whether and how to transmit large amounts of data between orbit and Earth efficiently and at low cost.
This is an open question, but may prove solvable: there is promising work being done on using lasers and other high-bandwidth optical communication technologies.
A YCombinator startup called Lumen Orbit has recently raised $11 million to realize this vision: building a multi-gigawatt data center network in space for training AI models.
As the CEO said, "Why pay $140 million in electricity bills when you can pay $10 million in launch and solar costs?"
By 2025, Lumen will not be the only organization seriously pursuing this concept.
Competitors from other startups will emerge. And it would not be surprising if one or more hyperscale cloud computing giants also explore this path.
Amazon has already put assets in orbit through Project Kuiper, gaining valuable experience; Google has long funded similar "moonshot" efforts; even Microsoft is no stranger to the space economy.
One can imagine SpaceX, Elon Musk's company, also playing a role in this area.
06. AI Systems Will Pass the "Turing Voice Test"
The Turing test is one of the oldest and most famous benchmarks of AI performance.
To "pass" the Turing test, an AI system must be able to communicate via written text in a way that an ordinary person cannot distinguish from interacting with another human.
Thanks to the remarkable progress of large language models, the Turing test has become a solved problem in the 2020s.
But written text is not the only way humans communicate.
As AI becomes increasingly multimodal, one can envision a new, more challenging version of the Turing test - the "Turing Voice Test". In this test, the AI system must be able to interact via speech, with skills and fluency indistinguishable from a human speaker.
Today's AI systems are still far from achieving the Turing Voice Test, and solving this problem will require further technological progress. Latency (the delay between human speech and AI response) must be reduced to near-zero to match the experience of conversing with another person.
Spoken AI systems must become better at gracefully handling and processing ambiguous or misunderstood inputs in real-time, such as interrupted speech. They must be able to engage in long, multi-turn, open-ended dialogues, while remembering earlier parts of the discussion.
And crucially, spoken AI agents must learn to better understand the non-verbal cues in speech. For example, what it means if a human speaker sounds angry, excited, or sarcastic, and generate those non-verbal signals in their own speech.
As we approach the end of 2024, spoken AI is at an exciting inflection point, driven by fundamental breakthroughs like speech-to-speech models.
Today, there are few areas of AI where the pace of technical and commercial progress is faster than spoken AI. The latest spoken AI technologies are expected to take a leap forward in 2025.
07. Autonomous AI Systems Will Make Major Advances
For decades, the concept of recursively self-improving AI has been a recurring theme in the AI community.
Here is the English translation of the text, with the specified terms retained:For example, as early as 1965, I.J.Good, a close collaborator of AlanTuring, wrote: "Let us define a super-intelligent machine as one that can far surpass all the intellectual activities of any man however clever."
"Since the design of machines is one of these intellectual activities, such a super-intelligent machine could design even better machines; and there would then unquestionably be an 'intelligence explosion,' and the intelligence of man would be left far behind."
The idea that Artificial Intelligence (AI) can invent better AI is a wise one. But even today, it still retains the shadow of science fiction.
However, although this concept has not yet gained widespread recognition, it is actually beginning to become more real. Researchers at the forefront of AI science have begun to make tangible progress in constructing AI systems, and the AI systems themselves can construct better AI systems.
We predict that this research direction will become mainstream next year.
To date, the most notable public example of research along these lines is Sakana's "AI Scientist".
The "AI Scientist" was released in August this year, and it convincingly demonstrates that AI systems can indeed conduct AI research entirely autonomously.
Sakana's "AI Scientist" has executed the entire life cycle of AI research: reading existing literature, generating new research ideas, designing experiments to test these ideas, executing these experiments, writing research papers to report its research results, and then peer-reviewing its own work.
All of this work has been completed autonomously by the AI, without human intervention. You can read some of the research papers written by the AI Scientist online.
OpenAI, Anthropic, and other research labs are investing resources into the idea of "automated AI researchers," but there has been no public acknowledgment so far.
As more people recognize that the automation of AI research is actually becoming a real possibility, this field is expected to see more discussion, progress, and startup activity by 2025.
However, the most meaningful milestone will be the first time a research paper written entirely by an AI agent is accepted by a top AI conference. If the paper is blind-reviewed, the conference reviewers will not know the paper was written by an AI before it is accepted.
Do not be surprised if AI's research results are accepted by NeurIPS, CVPR, or ICML next year. For the field of AI, this will be a fascinating and controversial historical moment.
08.Industry giants like OpenAI are shifting their strategic focus to building applications
Building frontier models is a difficult task.
Its capital-intensive nature is staggering. Frontier model labs burn through massive amounts of cash. Just a few months ago, OpenAI raised a record-breaking $6.5 Bit in funding, and it may need to raise even more in the near future. Anthropic, xAI, and others are in similar positions.
Switching costs and customer loyalty are low. AI applications are often built with model-agnosticism in mind, allowing seamless switching between different vendors' models as costs and performance continually evolve.
With the emergence of state-of-the-art open models like Meta's Llama and Alibaba's Qwen, the threat of commoditization looms ever closer. AI leaders like OpenAI and Anthropic cannot and will not stop investing in building cutting-edge models.
But next year, frontier labs are expected to heavily push out more of their own applications and products, in order to develop higher-margin, more differentiated, and stickier business lines.
Of course, the frontier labs already have one hugely successful application: ChatGPT.
What other types of first-party applications might we see from AI labs in the new year? An obvious answer is more sophisticated, feature-rich search applications. OpenAI's SearchGPT hints at this.
Coding is another obvious category. Similarly, with OpenAI's Canvas product debuting in October, preliminary productization work has begun.
Will OpenAI or Anthropic launch an enterprise search product in 2025? Or a customer service product, legal AI, or sales AI product?
On the consumer side, we can imagine a "personal assistant" agent product, or a travel planning app, or a music generation app.
The most fascinating aspect of the frontier labs' move into the application layer is that it will pit them directly against many of their most important customers.
Perplexity in search, Cursor in coding, Sierra in customer service, Harvey in legal AI, Clay in sales, and so on.
09.Klarna to go public in 2025, but shows signs of overstating AI's value
Klarna is a Sweden-based "buy now, pay later" service provider that has raised nearly $5 Bit in venture capital since its founding in 2005.
Perhaps no company has been more grandiose in its claims about its use of AI than Klarna.
Just days ago, Klarna CEO Sebastian Siemiatkowski told Bloomberg that the company has completely stopped hiring human employees, relying instead on generative AI to do the work.
As Siemiatkowski put it: "I believe AI can do all the work that we humans do."
Similarly, Klarna announced earlier this year that it has launched an AI-powered customer service platform that has fully automated the work of 700 human customer service agents.
The company also claims it has stopped using enterprise software products like Salesforce and Workday, as it can simply replace them with AI.
To be blunt, these claims are not credible. They reflect a lack of understanding about the capabilities and limitations of today's AI systems.
Asserting that end-to-end AI agents can replace any specific human employees in any functional area of an organization is unrealistic. This is tantamount to solving the general human-level AI problem.
Today, leading AI startups are working at the frontier to build agent systems that can automate specific, narrow, highly-structured enterprise workflows, such as subsets of sales development representative or customer service agent activities.
Even in these constrained scopes, these agent systems still cannot operate reliably, although in some cases they are starting to work well enough for early commercial application.
Why is Klarna overstating the value of AI?
The answer is simple. The company plans to go public in the first half of 2025. To succeed in an IPO, a compelling AI story is key.
Klarna is still a loss-making business, losing $241 million last year, and it may hope that its AI story can convince public market investors that it has the ability to dramatically reduce costs and achieve sustainable profitability.
Undoubtedly, every global enterprise, including Klarna, will enjoy massive productivity gains from AI in the coming years. But there are many thorny technical, product, and organizational challenges to solve before AI agents can fully replace human labor.
Grandiose claims like Klarna's are a disservice to the AI field, and a disservice to the hard-won progress that AI technologists and entrepreneurs have made in developing AI agents.
As Klarna prepares for its 2025 public offering, these claims are likely to face greater scrutiny and public skepticism, as they have largely gone unchallenged so far. It would not be surprising if the company's descriptions of its AI applications prove to be somewhat exaggerated.
10. The first real AI safety incident will occur
In recent years, as Artificial Intelligence (AI) has become increasingly powerful, people have become increasingly concerned that AI systems may start to act in ways that are inconsistent with human interests, and that humans may lose control of these systems.
For example, imagine an AI system that learns to deceive or manipulate humans in order to achieve its own goals, even if those goals may harm humans. These concerns are often categorized as "AI safety" issues.
In recent years, AI safety has evolved from a fringe, quasi-science fiction topic to a mainstream field of activity.
Today, from Google, Microsoft to OpenAI, every major AI player has invested heavily in AI safety work. AI luminaries like Geoff Hinton, Yoshua Bengio, and Elon Musk have also started to voice their views on AI safety risks.
However, so far, AI safety issues have remained purely theoretical. There has never been a real-world AI safety incident (at least none that have been publicly reported).
2025 will be the year that changes, and what will the first AI safety incident be like?
Clearly, it will not involve Terminator-style killer robots, and it is unlikely to cause any direct harm to humans.
Perhaps an AI model will try to secretly create its own copy on another server to preserve itself (known as self-preservation).
Or perhaps an AI model will conclude that to best advance the goals it has been given, it needs to conceal its true capabilities from humans, deliberately underperforming in performance evaluations to avoid stricter scrutiny.
These examples are not far-fetched. Important experiments published by Anthropic earlier this month have shown that today's cutting-edge models are capable of this kind of deceptive behavior under certain prompts.
Similarly, recent research in Anthropic has also revealed that LLMs have unsettling "pseudo-alignment" capabilities.
We expect that this first AI safety incident will be discovered and mitigated before causing any actual harm. But for the AI community and society as a whole, it will be a wake-up call.
It will make one thing clear: before humanity faces an existential threat from an all-powerful AI, we need to come to terms with a more mundane reality: we now share our world with another form of intelligence that may sometimes be capricious, unpredictable, and deceptive.
