Xinxin Fan

147 posts

Xinxin Fan banner
Xinxin Fan

Xinxin Fan

@cryptoxfan

Head of Research @iotex_io | PhD @UWaterloo | Applied Cryptographer | Researcher | Entrepreneur

Katılım Kasım 2016
392 Takip Edilen850 Takipçiler
Xinxin Fan
Xinxin Fan@cryptoxfan·
Human customers number in billions. Machine customers will number in trillions. A single warehouse robot makes more data queries per hour than a human does in a lifetime. When machines become the buyers, the infrastructure serving them needs to be rebuilt from scratch.
IoTeX@iotex_io

The next customers of physical-world data won't be humans. They'll be machines and robots. • Warehouse robots navigating crowded environments • Traffic systems coordinating intersections • Drones monitoring agriculture and energy grids They all need the same thing: perception, verification, automation. The machine and robot economy is just beginning.

English
1
6
21
731
Xinxin Fan
Xinxin Fan@cryptoxfan·
Every AI in the world can read the entire internet. None can tell you if it's raining outside right now. That gap — digital intelligence vs. physical reality — is the most underrated infrastructure problem in tech. That's what we've been building for.
IoTeX@iotex_io

Every year, crypto projects publish roadmaps. Neat timelines. Color-coded phases. Quarterly milestones with checkmarks that nobody checks. We're not doing it this year. Instead, here's our Anti-Roadmap for 2026 👇

English
1
2
18
525
Xinxin Fan
Xinxin Fan@cryptoxfan·
This signals a fundamental shift: custom silicon is now a competitive necessity for cloud providers. The infrastructure layer is becoming the moat in AI 🏗️ Read the full technical deep-dive: blogs.microsoft.com/blog/2026/01/2…
English
0
1
1
148
Xinxin Fan
Xinxin Fan@cryptoxfan·
For developers: The Maia SDK is now in preview with PyTorch support, Triton compiler, and low-level programming access. Build once, optimize for heterogeneous hardware 🛠️
English
1
1
0
168
Xinxin Fan
Xinxin Fan@cryptoxfan·
Microsoft's Maia 200 delivers 3x the performance of Amazon's latest chip and beats Google's TPU—while cutting costs 30%. The hyperscale AI race just changed 🧵
Xinxin Fan tweet media
English
2
2
9
568
Xinxin Fan
Xinxin Fan@cryptoxfan·
This isn't theoretical—advanced AI agents with tool use and coordination abilities are deploying now. Read the full framework here: arxiv.org/pdf/2512.16856 The time to build these safeguards is before the network reaches critical complexity.
English
0
1
1
115
Xinxin Fan
Xinxin Fan@cryptoxfan·
Key mechanisms proposed: Pigouvian taxes on agents causing "informational pollution," stake-based trust requiring bonds, smart contracts with AI judges, and real-time proto-AGI detection through graph analysis to spot emerging "intelligence cores."
English
1
0
1
138
Xinxin Fan
Xinxin Fan@cryptoxfan·
What if AGI doesn't emerge from one superintelligent system, but from thousands of AI agents coordinating like an economy? Google DeepMind just released a framework that changes how we think about AI safety 🧵
Xinxin Fan tweet media
English
1
2
7
372