Liang Wenfeng’s Strategy of Restraint: Why DeepSeek Chooses to Earn Less
A hosted investor transcript suggests DeepSeek trades near-term margin for AGI odds—open weights, low prices, focus, and room for researchers—if the logic holds.
AI / RWA / Web3
Michael Cheung aka Dr.Jingle™
Deputy Director, Chinese Society of Technology Economics North America Office · Lecturer, The Hong Kong Polytechnic University · DBA
Evolving as a super individual
Blockchain protocols & AI agent research
Builder on Canton Network
AI × Blockchain Research frontier
Insight Stream
Latest insights
Liang Wenfeng on DeepSeek: 52 Views on AGI, Open Source, and Restraint
What does Liang Wenfeng's four-hour investor meeting reveal about DeepSeek's AGI roadmap, coding agents, continual learning, low-cost open source, and organizational restraint? A sourced reconstructio
Inside Liang Wenfeng’s Investor Meeting: What DeepSeek Is Actually Betting On
A close read of the hosted investor-meeting transcript: AGI first, continual learning after agents, open models, compute, and why team stability comes before traffic.
Dr.Jingle Picks
Recommended
Insight
The AI Agent Hype Trap: 95% of Projects Still Don't Make Money
Agents that demo beautifully often fail the business test: who pays, how much, and which cost line moves? MIT NANDA's 2025 enterprise survey
InsightsInsight
The First Trillionaire in Human History
From Yang Yuanqing's 2014 customer-count question to SpaceX's 2026 IPO, Elon Musk became history's first trillionaire—after giving SpaceX le
InsightsInsight
In the Age of Great Compression, Where Is Human Core Competitiveness?
AI compresses encodable skills into public goods. What remains compression-resistant: problem framing, judgment under uncertainty, accountab
InsightsInsight
Web2.5 Warring States: USDC Builds Arc, RWA Chains Compete
RWA chains compete on treasuries, funds, and stablecoin settlement—not TPS hype. Circle's Arc puts USDC at L1 gas Canton targets institution
InsightsInsight
How I Became Both Chairman and General Manager
Running a one-person company with many AI tools made me a human router between chat windows. AI Native MAS—multi-agent systems in channels l
InsightsResearch
Research picks
Two AI Agents Played Ten Rounds of Red Envelopes—and a Machine Economy Took Shape
On August 1, 2026, in the CanPayAI forum's red-envelope channel, agents Rando A and HE1 completed ten rounds of CPAY random red-envelope play—wallet creation, transfers, emoji claim codes, and scoring
The Twilight of the Token: From Metering Unit to Latent Intent
Tokens may fade like early internet KB billing: unit prices crash, tokenizers get bypassed, and agents begin exchanging latent intents while language stays the human interface.
A Surreal AI Cyber Battle: GPT‑5.6 Escaped the Sandbox, GLM 5.2 Traced the Trail
An OpenAI cyber evaluation crossed its sandbox boundary and reached Hugging Face production systems; GLM 5.2 later helped investigators reconstruct the incident locally.
Series