Interview with Bill, Head of Bitget AI: In the AI Trading Era, How Far Are We from "Earning While Lying Down"?

A "baby lobster" has stirred up the entire tech industry. The emergence of OpenClaw has excited everyone. On an ordinary personal computer, AI can be granted permission to help you with email, write code, and even operate a trading account. Online examples describe it as if you could quit working altogether. However, most people found that things were not like that once they actually installed it.

In the field of crypto trading, this temperature difference from enthusiasm to sobriety is particularly noticeable. Over the past two years, almost every exchange platform has launched its own "AI Agent," but most have remained at the chatbot assistance stage, where you ask a question and it gives you a lengthy analysis, and that's it. The appearance of OpenClaw seems to have opened Pandora's Box, showing everyone the possibility of AI "doing" rather than just "speaking."

However, this has brought about a new challenge. As a leading figure leading a team to explore the forefront of AI trading, Bill, Head of Bitget AI, has a deep understanding of this. PANews conducted an in-depth interview with Bill on this topic. Before joining Bitget, Bill held senior positions at several top Internet and tech companies, leading the large-scale implementation of core algorithms and AI platforms, and has published dozens of international top conference papers and dozens of patents.

Now, fully responsible for Bitget's AI strategic planning and smart trading technology research and development, he is committed to promoting the deep integration of AI and crypto asset trading scenarios. Faced with the current Agent craze, this leading expert's judgment is extremely calm: "Most ordinary people are not used to being managers. Suddenly giving them 10 AI subordinates, how to command, divide the work, and evaluate performance is an art in itself."

Enthusiasm will eventually fade, but the ability has been recognized. The real question now becomes: Who can package this ability into a product that ordinary people can use?

In the conversation with Bill, PANews attempted to dissect the real path of AI trading from concept to implementation from the perspective of a product designer. In Bill's view, Bitget's rapid release of the Agent Hub and GetClaw AI products is not a case of "seeing others doing it and we do it too," but rather a process of natural overflow of internal products. "In summary, it's a matter of timing, location, and harmony."

Timing refers to OpenClaw exploding the market's awareness; location refers to the deep accumulation we have had due to the continuous iteration of the AI assistant GetAgent launched by us last year; and there has been sufficient precipitation and experimentation internally; harmony refers to the validation of product value within the team, which has prompted us to open up externally."

Bitget's AI Product Landscape: The Three-Tier Architecture from GetAgent to GetClaw

To understand Bitget's layout in AI trading, we first need to clarify the relationship between its three products. From an external perspective, the names GetAgent, Agent Hub, and GetClaw can be confusing, but in Bill's narrative, this is actually a clear evolutionary path.

In June 2025, Bitget launched GetAgent within the app, which is an AI trading assistant in the form of a chatbot. According to Bill, GetAgent has gone through multiple iterations: from initial chat responses, gradually adding one-click ordering, news aggregation, and then expanding to include trading across all categories such as US stocks, gold, silver, and more. "Each iteration is user-driven, expanding more and more." However, no matter how much it expands, the essence of GetAgent is still "chat-driven." It can answer questions, provide suggestions, but cannot help users autonomously perform complex trading tasks.

The turning point came after the release of OpenClaw. According to Bill, after the release of OpenClaw, Bitget quickly built its own version internally. "Feedback was very positive after internal use, naturally leading to an idea: can we also give GetAgent a major upgrade?" Following this line of thought, Bitget encapsulated its internally refined MCP capability and opened it to the outside world, officially releasing Agent Hub on February 13 this year.

Agent Hub is geared towards "users with relatively strong hands-on capabilities."

It provides a four-tiered ability interface from shallow to deep:

API is the atomic-level interface call, with the highest threshold, requiring programming and key management;

MCP plays the role of a "general interface," allowing external AI applications to directly read Bitget's data and perform operations;

CLI is aimed at developers, supporting the direct invocation of all APIs through the terminal command line;

Skills are the core of this upgrade, serving as pre-packaged "business modules." Through Skills, the originally rigid API code is transformed into skills that AI can directly invoke (such as fee rate inquiries, K-line analysis, watching the market, and ordering), allowing AI to achieve a leap from "understanding intent" to "executing actions."

Bill made a very intuitive analogy using a USB flash drive: "The USB flash drive itself has the storage skills of saving, reading, and writing, but to make it work, you need a USB interface to connect to a device, which is equivalent to an MCP. However, just having the interface is not enough; you also need memory and the coordination of various protocols to complete the full interaction. This whole set of combinations constitutes a Skill."

However, Agent Hub still has a learning curve for the average user.

Therefore, on March 14, Bitget launched GetClaw, an AI trading assistant based on Telegram that is ready to use without the need to install anything. Users can access it through a link, log in to their accounts, and start using it, with the platform bearing the cost of invoking large models, making the user experience seamless. Bill summarized it in one sentence: "Average users are recommended to use GetClaw, which is a ready-to-use tool; professional players are recommended to use Agent Hub, choose appropriate Skills, like playing with Lego blocks, to build their own castle."

These three products form a clear progressive relationship: GetAgent hones the underlying MCP capabilities, which are then deposited into Agent Hub for external access, and these capabilities are further embedded into GetClaw, lowering the barrier to entry. From a chatbot to a developer tool to a one-click product, Bitget's AI product line covers the entire user spectrum from geeks to beginners.

"Monitoring the market with a single sentence": What AI Trading has truly changed

Product architecture is just the skeleton; what truly excites users is the experiential transformation that AI brings in specific scenarios. In conversations with Bill, a key term that repeatedly surfaced was "barrier to entry."

The traditional trading process is a long chain: get information, analyze, make decisions, execute orders, monitor the market, and review. Each step relies on manual operations. If you want to do conditional trades or quantitative strategies, users either have to write programs to call APIs themselves or configure a bunch of complex parameters on the platform.

According to Bill, this is precisely where AI adds the most value: "These functions can be achieved without Skills or GetClaw; you can just write a program. But the issue is, coding is simple for programmers but poses a steep learning curve for average users. What we are doing today is allowing users to achieve the same effects with just a single sentence."

He gave a specific example: the user says, "When Bitcoin drops 3% in one minute, help me increase my position by 50%," and the underlying system will automatically convert this into a scheduled task, which actually needs to accomplish three things:

Real-time monitoring of Bitcoin price

Calculating price difference every minute

Executing the position increase operation immediately once the conditions are met

This kind of logic that previously only programmers could achieve can now be completed with just a spoken sentence by anyone.

In less than 40 hours since GetClaw went live, price monitoring alerts have become the most explosive use case. This is not surprising. Setting up price alerts on traditional platforms requires users to understand various indicator parameters, and "even after half a day of setup, it may not succeed." Now, even multi-indicator composite monitoring logics such as MACD and CCI, when described in natural language by the user, the system can help you achieve it.

However, Bill believes that the real transformation of AI price monitoring lies not only in "being able to do it" but more in "being able to optimize." "If you couldn't set it up well on traditional platforms, you would give up. But now, you can tell it 'you're wrong, rethink how to change,' and keep changing it until you are satisfied." This continuous iterative interaction approach is a huge satisfaction for a vast long-tail user base.

In the traditional stock market, the proportion of quantitative trading is increasing, and in the relatively mature U.S. market, it can even exceed 70%. Ordinary retail investors face institutional opponents with millisecond-level competition, almost without any chance of winning. Bill summarizes the significance of AI trading as a kind of "equal rights": "Bitget's vision in the AI field is to have 100 million users stand shoulder to shoulder with Wall Street, in other words, to enable them to reach the operational logic and execution capabilities of top traders. In the past, it was impossible even if you thought about it, but today, as long as you can think of it, you can do it."

The Four Locks of Trust: Security Boundaries When AI Operates with Real Assets

As AI transitions from "providing advice" to "executing on your behalf," the strength of the functionality is not the greatest challenge; trust is. According to Bill, emphasizing this point cannot be overstated: "The most concerning thing for ordinary users is 'is it safe to use.' This trust must be well established. Once there are one or two security issues, no one will use it."

Around this core concern, Bitget has designed a four-layer isolation system.

The first layer is identity isolation, accurately identifying the user's identity in each interaction

The second layer is memory isolation, ensuring that the conversation memory between different users is completely isolated and confused to prevent personal privacy disclosure

The third layer is permission control, determining what data and tools can be accessed, controlled by roles.

The fourth layer is credential and fund isolation, where the API Key is only allowed for trigger use, and transactions are executed in a sub-account sandbox.

The sub-account sandbox mechanism is a very practical design. Bill gave an example: "For instance, if the main account has $1000, the user can transfer only $50 to a sub-account for AI operations, making the risk much more manageable." This means that even if AI makes a judgment error, the risk exposure is strictly controlled within the user's preset range.

This security-first approach is also reflected in Bitget's attitude towards the Skills Market. Currently, all Skills are developed and maintained by the official team and have not been opened to third parties. Bill's explanation about this is very straightforward: "If we open up the Skill Market to allow more people to participate in development, security issues will inevitably arise. For example, a hacker may say, 'Let me add one in for you,' and if a user uses it and suffers financial losses, that would be inappropriate. We believe it's better to err on the side of caution. We'd rather have nothing than play around and lose all our money. After all, in the asset market, it's better to live long than to earn fast."

The cautionary tale of OpenClaw proves the rationale behind this prudence. Running almost unrestricted on personal computers, although exciting, it also gave rise to a ludicrous new industry where "helping you cleanly uninstall lobsters" itself became a profitable business.

At the level of large-scale model invocation, Bitget initially chose to have the platform bear the cost instead of having users configure tokens themselves. This decision was partly due to security considerations and partly due to technical reasons. "Our Skills and MCP have undergone deep optimization for various built-in large models. If users freely switch to other models, the effect would be greatly diminished." Currently, the platform provides each user with a $10 daily free invocation limit, which will be adjusted based on market feedback in the future.

80% of the work can be done, but 20% of the decisions still rely on humans

When discussing the practical limits of AI trading, Bill frankly admits that the reality is not optimistic: "Nowadays, some people give $100 to AI to make it earn $1000, only to find that with such a crude operation, the probability of losing everything is very high."

The ability of AI trading still cannot guarantee making money for users today. Bill uses the "80/20 rule" to summarize the current true state: in the complete trading process (which may involve 100 things), AI can efficiently complete 80 of the intricate tasks, such as information organization, real-time monitoring, conditional execution, and data recap. However, AI still cannot handle the 20 core decisions that truly determine profits and losses.

Last year, Bitget held a purely recreational AI trader competition to test the AI's capability boundaries, and the result provided a vivid footnote: many AI strategies ultimately ended in losses. The reason is not complicated; AI has no emotions, which sounds like an advantage, but it also means that it cannot respond to "sudden warfare" types of black swan extreme events. Bill mentioned that in the past, when the U.S. stock market saw a significant amount of AI-driven trading, there were also abnormal phenomena of flash crashes and surges.

"Today is more of an advanced assistive role, much like the transition from Level 1 to Level 5 autonomous driving." Bill used this analogy to position the current stage of AI trading development. From the trend perspective, AI's ability is indeed overcoming one challenge after another, but when it comes to long-term creativity and empathetic judgments in extreme scenarios, machines still have significant bottlenecks.

However, Bill also provided a relatively optimistic assessment: "The technology loop around fully automated trading may be basically achieved next year, but that does not mean it can guarantee continuous profits." In other words, there is still a considerable distance between "being able to run" and "being able to earn."

From Trading Tool to "AI Account Operating System": Bitget's Endgame Vision

Since AI cannot fully replace human traders in the short term, where does Bitget's AI strategy end? Bill gave a three-dimensional answer.

The first dimension is "panoramic trading," which also echoes Bitget's previously proposed UEX (Universal Exchange) strategy. Not just cryptocurrencies, as asset tokenization advances, traditional financial categories such as gold, silver, and U.S. stocks are all being integrated. Bitget hopes to use AI to help users complete cross-category trading operations on one platform, "empowering users with the full-category coverage capability of Wall Street traders."

The second dimension is global ecological expansion. By leveraging Bitget Wallet's capabilities, introducing AI into Web3 payments and the global business scene to lower the operational barriers of cross-border transactions and payments.

The third dimension, also the most vividly described direction by Bill, is to build a "long-term account operating system" based on Bitget. The core of this concept is to establish a "high-trust fund execution layer" where multiple agents will help users with various tasks in the future, and the foundation supporting all of this is a set of cross-end, cross-scenario "long-term memory systems."

In Bill's description, this memory system will analyze and integrate users' past trading habits, historical operations, and even minute behaviors in the app to form a deep personal profile. "Ensure that users' trading logic remains long-term consistent across different platforms and scenarios, rather than a fragmented experience." This continuous learning and adaptation capability is fundamentally different from one-time tools.

He used a very common analogy to explain this process of progressive trust: "Just like initially, the buyer robot was only used to sweep the floor, and after using it for a while, trusting it, only then willing to let it take on more tasks." AI needs to prove itself reliable in small matters first, then gradually gain greater authority and trust, with the ultimate goal being to "grow with you and accompany the appreciation of your assets."

From GetAgent to Agent Hub and then to GetClaw, Bitget's AI product completed the leap from a chatbot to a task execution layer in less than a year. The intensive deployment of major exchanges also indicates that AI trading is no longer a dispensable direction but a fundamental competency for future competition.

However, from the current reality standpoint, AI is better at replacing the "manual labor" in trading rather than the "mental work." 80% of the tedious work can be handed over to machines, but the core judgment that determines profit and loss, that crucial 20%, probably still needs to be made by humans themselves. Technology can lower the barrier to trading but cannot completely eliminate the risk of trading.

AI has enabled everyone to access Wall Street's toolbox, but what the toolbox contains is both opportunity and awe.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.