Micky Stonez.ton🥚💎 (@mickystonecrypt) - Twitter Profili

Sabitlenmiş Tweet

Micky Stonez.ton🥚💎@mickystonecrypt·22 Oca

GM $TON of believers There have been some subtle but important changes in the CoCoon documentation, and they say a lot about where this is all heading. One of the first things that stands out is the mention of a CoCoon chatbot, clearly marked as “coming soon.” That alone hints at native AI interaction being built directly into the CoCoon stack, not something added as an afterthought. Another major update is the introduction of a public API to client endpoints. This is a big shift for developers. Instead of setting up and maintaining expensive servers, they’ll be able to connect directly to CoCoon’s ready-made infrastructure. Less overhead, faster development, and a much smoother path to shipping real products. It’s also now very clear that no matter how you choose to interact with CoCoon, payments are done in TON. There are no alternative rails mentioned and no workarounds implied. TON is the fuel for the entire system, and that feels like a very intentional design decision. One interesting detail is what’s not there. The documentation makes no mention of Telegram celebrities or influencers at all. That’s quite curious, but also telling. It doesn’t prohibit their involvement, it just doesn’t give them any special treatment. CoCoon is being positioned as neutral infrastructure, not something built around hype or personalities. Overall, this feels less like a marketing push and more like quiet groundwork. AI interfaces, shared infrastructure, and TON-native payments are being lined up so builders can focus on creating, not wrestling with backend complexity. The direction is becoming clearer, and the pieces are starting to fall into place.

English

15

5

65

3.8K

Micky Stonez.ton🥚💎 retweetledi

Micky Stonez.ton🥚💎@mickystonecrypt·29 Nis

I think AGI will combine the power of pattern learning from neural networks and the power of reasoning/understanding of symbolic systems called neuro symbolic AI, which is the deterministic path to AGI, not the smart parrot (LLM) that we have today, because pure pattern learning (LLM) still struggles with systematic generalization, compositional reasoning, and reliable verification and the purpose of AGI is to be able to generalise across domain and solve the problem he haven't seen before without retraining. So if an AI can think logically, understand the world, and verify its own logic is mathematically sound, it can reliably generalize across many domains, solve genuinely new problems, and adapt without task specific retraining. So I expect AGI to come from some mix of neural learning, symbolic reasoning, world models, planning, tool use, and grounding all in one, which is the real AI Durov is talking about even though he avoids using the marketing buzz word of AGI.

Micky Stonez.ton🥚💎@mickystonecrypt

GM $TON of believers Still on what Pavel Durov said about his brother Nikolay building “real AI” and why he believes it goes far beyond current AI systems. According to @durov, today’s large language models (LLM) don’t actually think. They just process huge amounts of text and generate responses based on statistical patterns, what he calls a “consensus version” of information. It can sound convincing, but it isn’t necessarily true. The core problem, in his view, is that humans confuse fluent language with real intelligence. He contrasts this with what his brother Nikolay is building, an AI that doesn’t just generate text but can think logically and understand the world itself. Durov describes Nikolay as a “genius,” and also clarified that he hasn’t been involved in Telegram’s operations for many years. Instead, he has been focused on fundamental research, including work on an infinitely scalable blockchain architecture. On the broader impact, Durov believes AI will reshape the labor market significantly. He expects job losses across professions like law and medicine and says the transition will be difficult for experienced workers, potentially reducing their value. At the same time, he believes younger people will adapt more easily and that new categories of jobs will eventually emerge. The core message is simple. We are not just dealing with better AI tools, we are looking at a potential shift in how intelligence itself is defined and how entire job markets restructure around it.

English

1

7

460

Micky Stonez.ton🥚💎 retweetledi

Micky Stonez.ton🥚💎@mickystonecrypt·29 Nis

GM $TON of believers Still on what Pavel Durov said about his brother Nikolay building “real AI” and why he believes it goes far beyond current AI systems. According to @durov, today’s large language models (LLM) don’t actually think. They just process huge amounts of text and generate responses based on statistical patterns, what he calls a “consensus version” of information. It can sound convincing, but it isn’t necessarily true. The core problem, in his view, is that humans confuse fluent language with real intelligence. He contrasts this with what his brother Nikolay is building, an AI that doesn’t just generate text but can think logically and understand the world itself. Durov describes Nikolay as a “genius,” and also clarified that he hasn’t been involved in Telegram’s operations for many years. Instead, he has been focused on fundamental research, including work on an infinitely scalable blockchain architecture. On the broader impact, Durov believes AI will reshape the labor market significantly. He expects job losses across professions like law and medicine and says the transition will be difficult for experienced workers, potentially reducing their value. At the same time, he believes younger people will adapt more easily and that new categories of jobs will eventually emerge. The core message is simple. We are not just dealing with better AI tools, we are looking at a potential shift in how intelligence itself is defined and how entire job markets restructure around it.

Micky Stonez.ton🥚💎@mickystonecrypt

GM $TON of believers In June 2025 Pavel @durov said in an interview with Le Point: “My brother Nikolai Durov is currently working on a real AI an AI that can think logically and understand the world.” Durov is describing an AI that can think logically, understand the world and verify its own logic is mathematically sound before producing output. This is known as Neuro symbolic AI. This signals a massive shift from smart parrots (LLMs/narrow AI) to a logical thinker (neuro symbolic AI or AGI). Instead of pattern based prediction it moves toward structured reasoning world understanding and mathematical self verification before output. This AI model is expected to be open source and deeply integrated into COCOON(confidential compute open network). Pavel Durov’s description of Nikolai Durov’s work aligns with the core scientific definition of Artificial General Intelligence even if he avoids using the marketing buzzword. According to researchers Neuro symbolic AI is one of the only deterministic paths toward AGI.

English

8

2

26

1.8K

Micky Stonez.ton🥚💎@mickystonecrypt·16h

@ahmedxm01 Congrats 🎉

English

0

1

23