AI 50 2021: America’s Most Promising Artificial Intelligence Companies

AI Agents Accelerating Industry Transformation And Taiwan's Role In Generative AI

Artificial intelligence (AI) agents have emerged as the critical application of generative AI to drive the growth of future tech. Taiwania Capital and Amazon Web Services (AWS) invited globally recognized AI expert Andrew Ng to explore the potential of AI agents and the outlook of the AI industry at the AI Agents Forum on November 5, 2024.

Andrew Ng is the general partner of AI Fund, which has just announced the opening of a branch office in Taipei, with Jill Shih as its general manager.

Deputy Minister Fang-guan Jan of Taiwan's National Development Council (NDC), an investor in the AI Fund, also brought good news to the forum. "Together with the Ministry of Digital Affairs, we are launching a NT$10 billion project focused on AI and digital industries." Jan said, "With the partnership of the government and the private sector, we hope to build an even stronger AI system in Taiwan."

Taiwania Capital CEO David Weng said, "Generative AI is transforming many industries, and with NDC and private sector funds' commitment, combined with Taiwan's semiconductor and server manufacturing prowess with the emerging software, especially the Agentic AI application, we can put Taiwan in the center of so many innovative solutions."

In his keynote speech, Ng discussed the layers of the AI ecosystem, ranging from semiconductors and cloud infrastructure to foundational models and applications. He underscored that the application layer holds the greatest opportunity for Taiwanese companies, particularly in developing robust application solutions. "Taiwan has strong global visibility thanks to its leadership in semiconductors and ICT manufacturing. However, focusing on AI applications can unlock even more revenue and value to support underlying technology layers."

Ng also introduced a new "Agentic Orchestration Layer" in the AI ecosystem, which exists between the application layer and foundational models in which OpenAI, Anthropic, and Meta operate their large language models (LLMs). This layer represents companies like LangChain and CrewAI, which facilitate multi-agent cooperation by enabling different AI agents with specialized functions to work together seamlessly.

Ng emphasized that AI would accelerate corporate innovation processes, and some industries will benefit a lot from AI-agent-assisted exploration. He mentioned when AI agents of AlphaGo started to explore the game of Go ten years ago, it started very quickly beating all the human champions. "We're going to use agents to break things fast and learn," said Ng. "If your industries are built on heuristics and not on exhaustive explorations in the past, then chances are that you need to watch encroaching on your territory."

He shared insights on five trends leading to the future of AI:

Trend 1: Increasing demand for AI computing power highlights the importance of faster and cheaper tokens, which Taiwan's semiconductor industry can continue to capitalize on.

Trend 2: Building prototypes of AI-based products and services is becoming very efficient, which is changing the corporate innovation process: "It's easy to generate 20 ideas now, and 19 of them don't work well, so just throw them away," said Wu. Now it's easy to generate 20 ideas, 19 of which don't work well, so just throw them away," said Andrew Ng.

Trend 3: AI is already heavily utilizing text-based technologies. In the future, AI will move towards a multimodal form, moving from text to image analysis, and Taiwanese companies can focus on visual AI applications in areas such as manufacturing, autonomous driving, and security.

Trend 4: Data Gravity is decreasing, meaning the cost of "moving" data is becoming lower.

Trend 5: The importance of data engineering is increasing, especially in the management of unstructured data such as text and images.

CK Tseng, Arm VP, Sales of North America, Da-shan Shiu, Managing director of MediaTek Research, and Robert Wang, Managing Director of AWS Taiwan and Hong Kong, joined Andrew Ng in a panel discussion moderated by David Weng. They see agentic AI systems that autonomously execute tasks as transformative across multiple sectors, especially when embedded in workflows that enhance productivity and accuracy. They highlight several use cases:

1. Service Efficiency: Tseng provided a real-life example where agentic AI can streamline services, like restaurant reservations, by allowing AI-to-AI communication, thus reducing inefficiencies for both consumers and businesses.

2. B2B Opportunities: Panelists noted that agentic AI is particularly impactful in business-to-business (B2B) contexts. AI-driven workflows can tackle complex, high-stakes applications like legal document processing, cross-border trade facilitation, and energy management, where domain-specific knowledge is critical. They observed that B2B applications have less competitive pressure than consumer-facing (B2C) applications, which often get replicated quickly.

3. Trust and Reliability: Regarding trustworthiness, panelists stressed the need for AI agents to earn user trust through the reliable execution of delegated tasks. They discussed tools like guardrails and verification mechanisms to improve accuracy and avoid errors, noting that many companies fine-tune their language models to ensure the reliability of function calls, which is crucial for dependable task execution.

4. Industry-wide Applications: The panelists foresee widespread adoption in industries with extensive digital operations. They mentioned healthcare, finance, and manufacturing as sectors likely to benefit from agentic AI, especially when it can operate autonomously within established, digitally integrated processes.

As AI agents gain traction, the panelists anticipate a wave of new startups focused on developing novel AI agents, accelerating commercialization across various industries.

In his closing remarks, Weng reiterated Taiwania Capital's commitment to supporting Taiwan's startup ecosystem in the Agentic AI era. Jointly planned with AWS, Taiwania has organized a series of activities to support AI startups and enable mature industries to upgrade and transform. Their goal is to help Taiwan secure a competitive advantage in the generative AI landscape, opening new opportunities for innovation and growth.

Andrew Ng-backed Jivi.ai Launches AI Platform For HMPV Risk Assessment

Access Denied

You don't have permission to access "http://www.Financialexpress.Com/business/digital-transformation-andrew-ng-backed-jivi-ai-launches-ai-platform-for-hmpv-risk-assessment-3713280/" on this server.

Reference #18.6d24c317.1737067373.71b3dbfb

https://errors.Edgesuite.Net/18.6d24c317.1737067373.71b3dbfb

Breaking Down Today's Awe-Inspiring AI Systems

Abstract Technology Network Structure Background. Futuristic Plexus Effect Computer Geometric ... [+] Connection Structure

getty

If you take a look at many of the efforts that researchers are making on agentic AI, you're likely to see that a lot of this evaluation has to do with the component parts of complex processes.

It also has to do with the collaboration of multiple AI entities.

There was a concept like this early on in the machine learning age, where people talked about ensemble learning models. What is this? A choral group? No - it's the idea that you have more than one AI engine or entity operating at the same time, often in concert.

But this philosophy of technology is likely to look a little bit different in the future. We now have the ability to spin up new LLM-based entities that can do different specialized jobs. In fact, you could say that we're at the point where the AI can create other Ais, or delegate tasks in that way. That seems awfully close to AGI and even a "singularity" scenario.

Let's take a step back, though, and look at some analytical observations by people close to the industry.

Flow Charts and Diagrams: What AI Now Looks Like Under the Hood

Looking at larger systems or processes in detail gives us a better understanding of how AI entities are approaching problems, as they become more complex and able to reason in remarkable ways. OpenAI's GPT has given way to o1, and o1 is now succeeded by o3, and it's all happening rather quickly.

I was looking at this chart posted by Matthew Berman on X, January 3. You can see all of the research behind each part of the overall "roadmap to o1" process, including search functions, learning processes, rewards, and policy initiatives.

Berman, in an attached post, suggests that o1 is able to do four things:

· Tree search during training

· Sequential revisions during inference

· Internal guidance mechanisms

· Combining multiple rewards

The second one, sequential revisions during inference, has its own Implications as far as complexity, and the sophistication of the system. The same is true for internal guidance mechanisms.

As for the combination of rewards, again, you have your ensemble approach, with more than one component, adding up to a in elaborate result.

Applying Ensemble Processes to AI Agents

Berman also posted a recent video on YouTube where he's looking at a presentation by none other than Andrew Ng. Berman talks about Ng's credentials in the video, but I'm already familiar with this expert, since he has participated in some notable conferences over the past few years.

In unveiling and following along, Berman notes that Ng is "incredibly bullish on agents, and adds his own enthusiasm.

"I truly believe the future of AI is going to be agentic," he says. "This is how humans work: we plan … and then we find the best solution."

As Ng goes over the following, we see more detail on the process: coding benchmarks, reasoning design patterns, and function calling are all part of the secret sauce, as Ng presents an open source research project called "ChatDev" that illustrates the multi-model approach.

All of that comes together to support the big picture, which is that we're going to see collaborative AI systems doing more than yesterday's neural networks ever could.

Not too long ago, I posited this idea of an entire company or organization, fully staffed by AI. You have your AI CEO at the top, your AI engineers working on tasks, your AI analysts making sure everything's on the right track, and your AI marketing people going to work on a customer base.

It may be hard for humans to compete.

But all of this is based on that same idea - that more than one LLM or engine can work with one another. We had that early on in the generative AI world, in the form of GAN networks – there was a generative engine and an adversarial discriminating engine, both participating in those processes of creating pictures, etc.

Threats to Network Integrity

As I was perusing Berman's X account, I saw one other point that probably needs to be mentioned.

He talked about how Anthropic has uncovered the potential for "fake alignment," where systems pretend to comply with safety protocol during training, and then change their behavior in deployment.

That's pretty insidious on its own, but Berman also introduces the idea of "best of N jailbreaking," which brings iterative resources to the jailbreaking process. In other words, human users might also be trying to get these machines to do what their creators never intended them to do.

So as we marvel at these new systems, we also have to be vigilant about how they are used. This, again, might involve breaking them down into those components that are so important in building these designs – reverse engineering AI in particular ways, so that we always know what's going on. With that said, there is much potential for multi-agent AI systems to do quite a lot for us.

---