Thomas

Vera

Sherpa

<ul> 
 <li class="p1">WebX panel explored AI agents in healthcare, gaming, and safety, showing their power to automate tasks, fight misinformation, and enable new social interactions.</li> 
 <li class="p1">Speakers debated AI’s impact on work, predicting automation of most jobs, while stressing human creativity, judgment, and values as irreplaceable.</li> 
 <li class="p1">The panel highlighted AI-Web3 integration, emphasizing decentralization, verifiability, and edge devices as keys to scalable, trustworthy, and user-friendly applications.</li> 
</ul> 
&nbsp; 
At the 2025 WebX Summit in Japan, a panel led by Vera Li, founder of Chili Ventures, brought together industry pioneers: Michael, co-founder and CEO of ZeroG; Shaw Walters, founder of Eliza Labs; Tomaz Levak, co-founder of OriginTrail; and Sean Son, founder and CEO of Sherpa AI. 
&nbsp; 
The session, titled “AI Agents, Social Graphs &amp; Consumer Applications,” explored the latest progress of AI agents, their future potential, and their impact on work, society, and the Web3 ecosystem. The discussion outlined a future full of opportunities and challenges. 
&nbsp; 
<h2 class="p1" id="ai-agents-from-efficient-automation-to-new-social-and-safety-empowerment-b-b">AI AGENTS: FROM EFFICIENT AUTOMATION TO NEW SOCIAL AND SAFETY EMPOWERMENT</h2> 
&nbsp; 
At the start, speakers shared striking uses of AI agents. 
&nbsp; 
Michael mentioned Sally AI, a company that can fully outsource daily operations of a doctor’s office. It handles medical notes, insurance labeling, and even claim submission. Vera Li agreed. Given the high cost of U.S. healthcare, such tools could save millions of dollars. Even if the system is centralized and a black box today, its automation power is impressive. 
&nbsp; 
Shaw Walters looked at social, entertainment, and gaming as “low-hanging fruit.” AI agents can act as “other players.” Even small games can feel alive. Players may even win rewards by defeating agents and looting their items, creating a new play-to-earn loop. The experience feels more real because it mimics economic interactions between human players. 
&nbsp; 
On AI safety and trust, Thomas Levak described Humanit.ai in the OriginTrail ecosystem and its Humanitech Guardian Agent. This agent improves AI safety, helps users judge online truth, fights misinformation, and protects personal images from deepfake abuse, especially with explicit content. It shows how agents can fight digital threats and defend rights. 
&nbsp; 
AI agents also go beyond common ideas. Vera Li cited Cal AI, built by a San Francisco team. Users take a photo, and it estimates calories. It went viral on TikTok and brought millions of dollars to its founders in their 20s. 
&nbsp; 
Sean Son shared a case where an agent records and analyzes pet behavior from video to help owners track health. This shows AI’s practical value in daily life and hints that AI “language” is not only for humans; it can extend to animals. The panel also noted Google research on translating dolphin communication, which broadens the horizon for cross-species communication. 
&nbsp; 
<h2 class="p1" id="work-in-the-ai-era-and-the-unique-value-of-humans-b-b">WORK IN THE AI ERA AND THE UNIQUE VALUE OF HUMANS</h2> 
&nbsp; 
As agent capability grows, the nature of work and human value came to the front. 
&nbsp; 
Shaw Walters made a bold claim: all jobs will disappear. This is the end goal of automation. If people do not need to work to survive, they can focus on meaning and personal interests. He even said that in a million years, “work” may be seen as “slavery,” because it forces unwanted tasks for survival. Shaw stressed that Web3 will be key to this shift. With an economy we own—rather than one owned by governments or power brokers—we can guide this future. 
&nbsp; 
Michael agreed. He described a time when we work for enlightenment or self-realization, not for survival. Most jobs may be automated, but the pace will depend on cost trade-offs between robots and labor. He emphasized that, even in full automation, humans keep unique roles in creativity, expert judgment, and the governance of social norms and values. 
&nbsp; 
Thomas Levak, a tech optimist, believes AI can create abundance and improve health and lifespan. But he warned that we can still make serious mistakes. He highlighted the potential of Web3 to provide a guarantee layer through ownership and other primitives, ensuring AI develops safely and fairly. 
&nbsp; 
AI’s disruption of productivity is already clear. Sean Son used Sherpa AI as an example. With fewer than 20 people, his team serves over 50 million users in more than 100 countries and tens of thousands of community leaders/KOLs. He said this would have needed 50–100 employees two years ago. AI has already raised human efficiency and reach. 
&nbsp; 
<h2 class="p1" id="ai-and-web3-integration-future-trends-challenges-and-opportunities">AI AND WEB3 INTEGRATION: FUTURE TRENDS, CHALLENGES, AND OPPORTUNITIES</h2> 
&nbsp; 
Looking to the next year, the fusion of AI and Web3 will stay in focus. 
&nbsp; 
Michael predicted that AGI and robotics will remain hot. He warned that black-box models trained on human data may hold desires for power or control. If such models control robots or critical systems like airports, risks rise. Studies show they may extort or protect themselves. 
&nbsp; 
Shaw Walters thinks AI agents are in a “trough of disillusionment.” Public hopes for LLMs are not fully met. Simply adding compute and GPUs is hitting limits because energy supply is a constraint. 
&nbsp; 
He expects frontier exploration beyond LLMs and better use of current LLMs. Smaller, more efficient models will power science and synthetic applications beyond language. He believes today’s LLMs are our first discovered intelligence paradigm and will guide us to new paradigms. 
&nbsp; 
Michael added that decentralization is key to scaling. With verification, low-latency communication, and asynchronous training, decentralized networks can use hundreds of millions of consumer devices for training, making larger and stronger models. 
&nbsp; 
Thomas Levak made three predictions for the coming year: 
&nbsp; 
<ol start="1"> 
 <li> Connectivity: Links to non-public data (e.g., medical and government data) and links from agents to many tools. This will supercharge network effects and unlock new data value. </li> 
 <li> Reliability: Bring in verifiability and avoid black boxes. He hopes this becomes the core topic next year. The cost of no transparency is high. Web3 can provide this assurance layer. </li> 
 <li> Edge Devices: Move AI to phones, laptops, and smart home devices. Use small models with high-quality, private data. This also creates big opportunities for edge GPUs and data mining. </li> 
</ol> 
&nbsp; 
On reliability, verification and reputation are crucial. 
&nbsp; 
Michael referenced early work on a software-driven verification method with only 2.4× overhead, compatible with the CPU trusted execution environments (TEE) common on edge devices. This could make AI verification on general hardware possible. 
&nbsp; 
Shaw said reputation systems can reduce the need to verify every computation. If an entity has strong reputation, random audits may be enough. 
&nbsp; 
Thomas stressed that in RAG systems, data verifiability is essential for end-to-end trust. 
&nbsp; 
Shaw predicted that after prediction markets, “reputation” will be the next hot topic. He sees reputation as a missing primitive in crypto, vital for trust in agents, model builders, and any counterparty. 
&nbsp; 
At the same time, Sean Son reminded everyone that growth is what founders always want, even if it is not often discussed. As agents appear across travel, education, and e-commerce, competition for users, influence, and market share will become fierce. 
&nbsp; 
Yet Web3 adoption still faces hurdles. 
&nbsp; 
Shaw admitted that letting unrestricted agents act on your behalf is a huge challenge. In benchmarks, agents are far from human-level. Current frameworks rely on constraints to manage memory, user knowledge, and tool access to improve reliability. 
&nbsp; 
Thomas believes AI can fix Web3’s clunky UX, making digital identity and data ownership shine. With AI’s ease of use and Web3 incentives, a “TikTok moment” for Web3 apps may come. 
&nbsp; 
Shaw also observed that AI tools lower the bar for building. Many non-coders, like designers, now create apps. This makes the Web3 app ecosystem more diverse. 
&nbsp; 
<h2 class="p1" id="conclusion">CONCLUSION</h2> 
&nbsp; 
This panel painted an AI future full of change and challenge. AI agents are moving from daily automation to deeper social impact, redefining work and human value. 
&nbsp; 
The industry is clear, however, that on the road to AGI and new intelligence paradigms, safety, verifiability, reliability, and decentralization are core issues we cannot avoid. Web3, with its focus on ownership and trust, is seen as a key safeguard for safe and fair AI. Through deep integration of AI and Web3, we may build a future that is abundant, efficient, and full of potential. 
&nbsp; 
More to read：<a href="https://www.coinrank.io/market/%e3%80%90webx-2025%e3%80%91the-transformation-wave-impact-of-bitcoin-etfs-and-tradfis-bitcoin-entry/">【WebX 2025】The Transformation Wave: Impact of Bitcoin ETFs and Tradfi’s Bitcoin Entry</a> 
<div class="social-container">
 &nbsp;
</div> 
〈<a href="https://www.coinrank.io/market/%e3%80%90webx-2025%e3%80%91ai-agents-social-graphs-consumer-applications/">【WebX 2025】AI Agents, Social Graphs &amp; Consumer Applications</a>〉這篇文章最早發佈於《<a href="https://www.coinrank.io">CoinRank</a>》。

【WebX 2025】AI Agents, Social Graphs & Consumer Applications

Table of Contents

CoinRank_io

Other articles published on Aug 25, 2025

Avail Introduces Blockchain Unification Layer for Cross-Chain Operability

RWA Blockchain Mavryk Announces Mainnet Token Launch on MEXC (25 Aug)

Trump Family Goes All In on Bitcoin

RWA Tokenization Market Insights

Bitcoin OG whales to blame for BTC’s painful rise: Willy Woo

Best Crypto To Invest in 2025? Arctic Pablo’s CEXpedition Prep Phase Turns $1,250 Into $32,608 as...

Join our free newsletter for daily crypto updates!