The key is building a trust architecture for AI now. Rogue entities could destabilize both the U.S. and China before they ever fight a war. Source: @nytimes @tomfriedman @BrookingsInst.#AI #USChina #Geopolitics.

A single rogue AI or a critical system failure could destabilize both superpowers long before a traditional conflict. This shared vulnerability to digital chaos makes building a trust architecture for AI a non-negotiable, top-priority security issue. 🧵 3/4.

What do you want to know?.

Unlike past technologies, AI will be ubiquitous, seeping into everything from consumer devices (watches, cars) to critical infrastructure (medical devices, power grids). This pervasive integration means a vulnerability anywhere can become a threat everywhere. 🧵 2/4.

The AI revolution presents a paradox for geopolitics. While driving fierce competition between the U.S. and China, its unique attributes will simultaneously force them into deep, unprecedented cooperation for mutual survival. 🧵 1/4.

These releases spark debate on how to properly evaluate AI, with experts pushing for practical, domain-specific tests. The trend highlights China's growing leadership in the open model space. Source: @smol_ai @swyx @karpathy.#AI #OpenSource 🧵 4/4.

Alibaba has also released Qwen 3 Max, a colossal 1 trillion parameter model. It is designed for complex, agent-like behavior and is already accessible via platforms like Qwen Chat and the Alibaba Cloud API. 🧵 3/4.

Moonshot AI's Kimi K2-0905 model has been updated, doubling its context length to an impressive 256k tokens. This allows it to process vastly larger amounts of information and improves its coding and tool-calling abilities. 🧵 2/4.

The open-source AI race is accelerating. Two massive new models, Kimi K2-0905 and Qwen3-Max, have just been released, signaling major advances in AI capability and scale, particularly from China. 🧵 1/4.

Teddy Roosevelt: "Do what you can with what you have where you're at.". This is the ultimate AI strategy. The real revolution isn't in the next breakthrough. It's in the massive gap between what's possible today and what we're actually doing. Build with purpose:.

Why do 95% of AI investments fail? The 3 P's Problem:. People: Using AI to summarize emails instead of thinking differently about work.Process: Retrofitting new tools into old workflows.Politics: Nobody knows who owns the AI strategy. Technology isn't the bottleneck. 2/3.

We're obsessed with "what's next?" in AI while 95% of organizations can't use what exists today. The "what's next" fallacy is killing real progress. Even if AI development stopped tomorrow, most companies would need YEARS to catch up with current capabilities. Stop waiting.

RT @ollama: Meet Google's new best small embedding model, EmbeddingGemma . It's a 300M embedding model made for retrieval augmented genera….

RT @gdb: Our custom LLM, gpt-4b micro, has helped achieve an advance in biology. It designed novel variants of the Nobel-winning Yamanaka….

Research Paper Summary. Title: rStar2-Agent: Agentic Reasoning Technical Report.Authors: Ning Shang et al. Link: Code/Data: Summary:.The paper introduces rStar2-Agent, a 14B parameter model that achieves state-of-the-art.

Stop chasing AI hype and start building a disciplined strategy. Learn how to avoid the experimentation trap from @HBR, based on findings from @mitmedialab and insights from @Gartner_inc. Source: #AI #BusinessStrategy #Innovation 🧵 4/4.

The solution is disciplined experimentation. Focus on high-intensity, high-frequency customer problems. Run low-cost pilots designed for iteration and empower small, effective teams to scale what works. ⚙️ 🧵 3/4.

The 'AI experimentation trap' is real. Instead of solving core customer problems, firms get lost in hype-driven distractions. This leads to isolated projects with no path to scale or measurable impact on the business. 🧵 2/4.

95% of generative AI investments have produced zero returns, reports the MIT Media Lab. This isn't an AI failure—it's a strategy failure. Many leaders are funding scattered pilots that don't connect to real business value, repeating past mistakes. 🧵 1/4.

The shift from "AI as tool" to "AI as partner" is happening faster than expected.From detecting cancer in blood to discovering new scientific principles autonomously. . What breakthrough surprised you most? 3/3 🧵.

What's fascinating: we're seeing convergence. Medical AI achieving clinical reliability.Scientific AI becoming autonomous.Reasoning AI becoming efficient.Memory systems getting more sophisticated. These aren't separate trends - they're building on each other 2/3 🧵.