The masterclass in hijacking the Google speakerโs stage, performed by @lunardragon420 at ETHPrague ๐๐
Brilliantly articulated the โbeautiesโ and benefits of CBDCs, centralized AI, and privacy-less networks.
This is 100% right. But especially in light of the recent AAVE and ZRO incidents, this whole area needs much stronger, more radical focus on SecOps and new insurance primitives. A single breach must not be able to push the entire DeFi sector to the edge of death.
Which means more privacy, more resilience and less dependence on centralized infrastructure.
You should consider such an initiative in your AI strategy.
AI is clearly here to stay and will reshape almost everything. But one topic that is still underestimated is - sovereignty.
And the world is moving toward a serious AI dependency trap:
The company is shifted beyond stablecoins into AI, peer-to-peer infrastructure and other emerging technologies. Their on-device AI through the QVAC framework allows AI to run directly on the device (even modern smartphones!) itself instead of sending all data to external servers.
Also recent geopolitical developments have shown that the world is becoming more fragmented and dependence on foreign infrastructure became a strategic weakness.
>> That is why I would pay close attention to what https://t.co/0PfZuvVwEn is doing <<
If those systems go offline and doctors have become too dependent on them over time, diagnoses may simply be delayed until the AI systems come back online. This creates a severe inability to execute without such systems.
If critical services become heavily dependent on AI, even a temporary outage can have serious consequences. Imagine hospitals relying on AI systems to analyze scans and identify illnesses.
It is very similar to what happened with cloud computing where we have only AWS, Google and Microsoft and couple of small players. Such a centralization only becomes visible when we experience global outages.
Remember Cloudflare and CrowdStrike? There is even more!
B) The second level is the compute infrastructure centralization. Even if you build your own AI applications, you still rely on someone elseโs compute, GPUs, cloud providers and data centers.
A) Most of the world consume AI from a very small number of AI providers such as OpenAI, Anthropic and Google. This creates the first level of vendor lock-in.