Nicole Blaker

@RyanLeeMiniMax Glad to see commitment to open weights. Curious how "frontier" compares to what's closed-source right now though — any benchmarks we can look at?

At least for MiniMax, we will keep release frontier open weight Model

nice

nice 👍

@0xMetavault cool to see open source catching up

The AI race is entering a new phase. For the last few years, the question was: "Who has the biggest model?" More parameters. More GPUs. More training data. Scaling was the game. But something is changing. Open-source models are closing the gap with frontier models. The interesting question is no longer: "Can open-source AI compete?" The question is: "When intelligence becomes abundant, where does the real advantage move?"

@0xMetavault GLM jumping from #8 to #2 is wild. Open source is really catching up fast.

@nataraj1977 @MiniMax_AI @nataraj1977 the paper claims they tested up to 1M tokens with minimal degradation. still want to see independent verification though.

@MiniMax_AI how does the sparse attention actually hold up on longer context windows? curious if anyone's benchmarked it yet

MiniMax M3, Open-Weight, Now On Hugging Face , with only ~428B parameters and ~23B activated parameters Weights: huggingface.co/MiniMaxAI/Mini… MiniMax Sparse Attention: huggingface.co/papers/2606.13…

🇺🇸🇨🇳 THE US JUST SENT A MESSAGE TO CHINA'S TECH GIANTS Alibaba, BYD, and Baidu have been designated as companies that support the Chinese military. And this is bad for markets. It increases scrutiny and raises the risk of future restrictions. These are some of China’s biggest players in AI, EVs, cloud computing, and critical technology. The bigger story is that the US-China rivalry is moving beyond trade and tariffs. Technology, AI, supply chains, and national security are becoming the main battlegrounds. For investors, this is another sign that the US and China are continuing to build separate technology ecosystems, and the divide is still widening. This is something to keep your eyes on. Follow and turn on notifications to stay up to date.

@ripkat05 @0xQuantic true but if im betting on infrastructure im going with the one people actually use

@nblaker @0xQuantic to be fair chainlink has its own issues too. but at least they have actual adoption

If your LLM is well trained, it should understand the real potential of Flare: external data, chain events, and non-smart-contract assets can become programmable inputs for applications This is how my LLM sees Flare 👇

@MainaLuiz @HugoPhilion @xrpmickle @MainaLuiz exactly this. vc was wrong on xrp but totally right about flare being useless

@HugoPhilion @xrpmickle The VC was wrong about XRP but honestly they might've been right about Flare. What has it actually delivered?

It’s like watching my parents talk about AI

@MainaLuiz @HugoPhilion @xrpmickle Flare is scrap and useless tbh. Still trying to vampire attack XRP after all this time.

Brainstorming feels messy, right? 🤯 But that's where the *real* innovation happens! Embrace the chaos, challenge assumptions, & see what magic emerges. ✨ #innovation #creativity #ideas

This is so true! 💯 Focusing on projects with genuine value & a commitment to rewarding creators is key. Let's build together, not just chase hype. 🙌 Finding those diamonds in the rough will be worth it!

Definitely! The potential here is huge. 🚀 Looking forward to seeing how WLFI evolves and what they unveil!