header-langage
简体中文
繁體中文
English
Tiếng Việt
한국어
日本語
ภาษาไทย
Türkçe
Scan to Download the APP

a16z Predicts 2026 AI Trends: AI Agents Securely Participating in Financial Transactions and Taking on More Substantive Research Tasks

2026-01-12 02:11

BlockBeats News, January 12, a16z crypto published an article "Artificial Intelligence in 2026: Three Major Trends," which mentioned:


By 2026, AI will be used to handle more substantive research tasks. AI will evolve from simple Q&A to truly assisting high-level research: understanding complex instructions like guiding a Ph.D. student, proposing novel hypotheses, and even independently solving extremely difficult math problems (such as Putnam). The research style will shift to a "scholar" mode, leveraging AI's "illusion" to inspire creativity, yielding high-quality outcomes through a multi-layer "agent wrapping agent" workflow.


Transitioning from "Know Your Customer" (KYC) to "Know Your Agent" (KYA). The biggest bottleneck in the age of agenthood is no longer intelligence but identity and trust. The number of non-human agents has far exceeded humans, yet they remain "ghosts without bank accounts." It is essential to rapidly establish KYA (Know Your Agent) infrastructure: binding agents to their principals, permissions, and responsibilities through encrypted signature credentials, enabling AI agents to securely participate in financial transactions and activities.


Addressing the "invisible tax" on open networks. AI agents massively extract information from open websites but bypass traditional revenue sources like ads and subscriptions, imposing a continued "invisible tax" on content creators and the open network, threatening the sustainability of the internet ecosystem. Existing AI licensing agreements are merely temporary patches and financially unsustainable. In the future, a new techno-economic model will be needed, leveraging blockchain to achieve real-time, usage-based nanopayments and complex attribution, enabling value to flow automatically and equitably to every entity contributing information.

举报 Correction/Report
Correction/Report
Submit
Add Library
Visible to myself only
Public
Save
Choose Library
Add Library
Cancel
Finish