Local vs. local decision models. One bay warehouse rule, reworded 11 ways. “A bay is eligible only if its handling is frozen.” “Bays without frozen handling are ineligible.” THIS/THAT 1.2 picked the right bay 93% of the time. Laya: 13%. Somewhere, ice cream is melting.

Our CEO @0x7SUN told Korea IT Times “For many critical enterprise tasks, generating open-ended text is an unnecessary cost and performance bottleneck.” “Specialized 2B-parameter models running locally can outperform massive cloud LLMs in speed, accuracy, and cost-efficiency.”

Proud for @DCGco to have backed @flock_io early - Qwen just spotlighted their open-source THIS/THAT model: 87.8% on complex decision tasks (up from 40.6% in v1.0), beating GPT-5.6 and Claude Opus 5, running locally in ~31ms. Built on Qwen3.5, trained on @alibaba_cloud . 🚀

Why does AI training even need blockchain? What are the real use cases for privacy-preserving and decentralised AI? Our Head of EMEA, @TiffanyWang98, takes on both questions, and more on the convergence of AI and blockchain with @0xmauricio on @blockdropspod. Out today 🎙️

We post-trained @Alibaba_Qwen's Qwen3.5 2B to build THIS / THAT Model 1.2, our open-source decision model, specialised for decision-making, routing, classification, and scoring. It scores higher than every hosted model we tested across 1,710 decisions and responds in 31.4 ms. That’s around 32x faster than Claude Opus 5. Training specialised models for specific business needs will become the norm.

Korea’s AI wave is getting hard to ignore. FLock 🤝 @Official_Upbit | Liquidity Support on the Final Day of KBW 2026. Great to be at @kbwofficial 2026 meeting builders, partners and friends pushing decentralised AI, crypto and blockchain forward. Until next KBW 🇰🇷

THIS / THAT model 1.1 is already here. Nearly 2x improvement on complex-decision test, ahead of Kimi, GLM & DeepSeek. Same 1.88B parameters. Same ~31ms. Build. Ship. Repeat. trained by FLock

My Model Token, Made in FOMO She said she wouldn’t FOMO, she said she wouldn’t FOMO. She lied, she lied, she lied. Her model token? Made in FOMO. Her model token? Made in FOMO.

The AI race is moving from compute to data. Much of the world’s most valuable data sits inside institutions, private companies, and governments, where it can’t simply be handed over to centralised AI providers. In a written interview with Korea IT Times, our founder and CEO Jiahao Sun (@0x7sun) explains how federated learning can unlock this data for AI training while keeping raw data local and how blockchain can add accountability and fairness to decentralised training. From healthcare in the UK to sovereign AI in Sarawak and public-sector innovation with UNDP in the Dominican Republic, this decentralized, privacy-preserving approach to model training is already being put to work. Many thanks to Chief Editor at Korea IT Times, Monica Younsoo Chung, for inviting us to contribute to the series. Read the full interview ↓

Your team knows how your business works. How do you build AI that does, too? At the builders' breakfast today with @UseCorgi, @fotor_com, and BuildHer Labs, our Head of EMEA, @TiffanyWang98, shared how businesses can build private, custom AI around their own knowledge and workflows. Your documents and policies give the model context. Fine-tuning helps it follow your processes and communicate in your brand’s voice. But these materials often contain sensitive information that needs to stay within your organisation. FLock provides the infrastructure for businesses to train and fine-tune domain-specific models through federated learning without exposing their raw data, and deploy them on infrastructure they control. A router connects each request to the appropriate model and applies rules on information access and responses. Smaller, custom models can handle routine work, while larger ones support more complex reasoning. The end goal is to build AI capabilities your organisation owns, shaped around real workflows and the people who use them. It was an insightful meetup where we discussed with fellow builders what enterprise AI solutions look like in practice. Thanks to @claudiaamaggi, Nora Chen, and Shreya Choudhuri for bringing us together, and to everyone who joined us at Corgi Cafe!

Governments worldwide are exploring how to adopt frontier technologies responsibly and securely. Today, under an executed framework agreement, we’re strengthening our relationship with @UNDP to advance public sector digital innovation in the Dominican Republic, leveraging distributed systems, privacy-preserving technologies, and AI applications. More details👇

It's never easy for a startup to sign an official framework agreement with a UN agency - the diligence, the alignment, the patience it takes to get there. But we've done it, and this marks our first step into sovereign AI and frontier tech collaboration in the LATAM region. We'll keep expanding on this, and we'll keep building. 🇺🇳🇩🇴

Governments worldwide are exploring how to adopt frontier technologies responsibly and securely. Today, under an executed framework agreement, we’re strengthening our relationship with @UNDP to advance public sector digital innovation in the Dominican Republic, leveraging distributed systems, privacy-preserving technologies, and AI applications. More details👇