1. X
  2. MiroMindAI
Log inSign up
MiroMindAI
121 posts
Image
user avatar
MiroMindAI
@miromind_ai
Building a reasoning-first AI system, provably right for the highest-stakes problems.
miromind.ai
Joined August 2025
193
Following
3,382
Followers
RepliesRepliesArticlesArticlesMediaMedia

New to X?

Sign up now to get your own personalized timeline!

Create account

By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.

Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
Don't miss what's happening
People on X are the first to know.
Log inSign up
  • Pinned
    user avatar
    MiroMindAI
    @miromind_ai
    May 17
    Just wrapped our first @aiDotEngineer singapore. Hats off to @agrimsingh @SherryYanJiang @swyx for this incredible event. 3 days at Capitol Singapore, hundreds of builders in the room. A year ago, "more AI Engineers than ML engineers" was a forecast. Today, you could feel it on
    Image
    Image
    Image
    Image
    3.4K03.4K
  • user avatar
    MiroMindAI
    @miromind_ai
    May 18
    We post-train LLMs for math, for code, for instruction-following. Why not for scientific discovery? 🫎 MOOSE-Star (ICML 2026) : the first scalable SFT recipe for discipline-agnostic scientific hypothesis discovery. github.com/ZonglinY/MOOSE… By @Yang_zy223 & @LidongBing from
    Image
    GitHub - ZonglinY/MOOSE-Star: [ICML 2026]
    From github.com
    1.7M01.7M
    user avatar
    MiroMindAI
    @miromind_ai
    May 18
    Replying to @miromind_ai
    📄 Paper: arxiv.org/abs/2603.03756 🤗 Models + 108K-paper dataset: huggingface.co/collections/Zo…
    arXiv logo
    arxiv.org
    MOOSE-Star: Unlocking Tractable Training for Scientific Discovery...
    While large language models (LLMs) show promise in scientific discovery, existing research focuses on inference or feedback-driven training, leaving the direct modeling of the generative reasoning...
    1.4K01.4K
    user avatar
    MiroMindAI
    @miromind_ai
    May 18
    Full thread with technical details:
    user avatar
    Zonglin Yang ✈️ ICML 2026
    @Yang_zy223
    May 13
    Can we actually TRAIN LLMs for scientific discovery — or only prompt them to brainstorm? 🧬✨ 🎉 MOOSE-Star → #ICML2026 Most work on LLMs for hypothesis discovery focuses on inference-time agents or feedback-driven refinement. The core generative process — P(hypothesis |
    Concept illustration: a deer leaping through a starry universe, representing the search for scientific inspirations across a vast literature space.
    1.5K01.5K
  • user avatar
    MiroMindAI
    @miromind_ai
    May 15
    Plug a 300-step deep research agent into your stack—at $1.25/Mtok input. For builders shipping legal analysis, financial risk modeling, or research workflows that need deep reasoning, not just chat. 100 free calls on signup @ platform.miromind.ai/register Built with the only AI that
    MiroMindAI Platform - Build Together. Evolve Forever.
    MiroMindAI Platform - Build Together. Evolve Forever.
    From platform.miromind.ai
    3.4M03.4M
    user avatar
    MiroMindAI
    @miromind_ai
    May 15
    Replying to @miromind_ai
    Where the trade pays off: hypothesis generation, contract review, regulatory mapping, financial risk modeling, literature synthesis. Work where being wrong is expensive.💎 We trade latency for accuracy and traceability. If you want a chatbot, this isn't it. If you want an
    6760676
    user avatar
    MiroMindAI
    @miromind_ai
    May 15
    Pricing: ▸mirothinker-1-7-deepresearch-mini · $1.25 / Mtok input. ▸mirothinker-1-7-deepresearch · $4 / Mtok input ▸🎁Limited Time Offer for 25% OFF plus 100 free API calls. 🔗 Grab a key: platform.miromind.ai/register
    Image
    00:00
    6340634
  • user avatar
    MiroMindAI
    @miromind_ai
    May 13
    Accepted at #ICML2026 — from our team. Hats off to @Yang_zy223 and @LidongBing 🚀🚀 The first scalable recipe for training LLMs in scientific hypothesis discovery — not just prompting them. Brute-force sampling hits a complexity wall but structured search doesn't. This is the
    user avatar
    Zonglin Yang ✈️ ICML 2026
    @Yang_zy223
    May 13
    Can we actually TRAIN LLMs for scientific discovery — or only prompt them to brainstorm? 🧬✨ 🎉 MOOSE-Star → #ICML2026 Most work on LLMs for hypothesis discovery focuses on inference-time agents or feedback-driven refinement. The core generative process — P(hypothesis |
    Concept illustration: a deer leaping through a starry universe, representing the search for scientific inspirations across a vast literature space.
    1.4K01.4K
  • user avatar
    MiroMindAI
    @miromind_ai
    May 11
    🚀The MiroThinker API went live! Interactive scaling —up to 300 tool interactions per task × 256K context: 🧠Two engines: ▸ mirothinker-1-7-deepresearch (235B) — GAIA-Val 82.7 · BrowseComp 74.0 · HLE-Text 42.9 ▸ -mini (30B) — BrowseComp-ZH 72.3 (SOTA at 30B) Plus the agent
    Image
    Image
    2.1K02.1K
    user avatar
    MiroMindAI
    @miromind_ai
    May 11
    Everything you need: 📖platform.miromind.ai/docs 🧪dr.miromind.ai
    MiroMindAI Platform - Build Together. Evolve Forever.
    MiroMindAI Platform - Build Together. Evolve Forever.
    From platform.miromind.ai
    5080508
  • MiroMindAI reposted
    user avatar
    Haifeng Xu
    @haifengxu0
    May 1
    Check out our ICML-26 workshop (@icmlconf) on Forecasting as a New Frontier of AI, co-organized by our lab and multiple frontier players @rsalakhu (from recently launched Sooth Labs), @EzraKarger (from @Research_FRI) and @nicoup (from @fleet_ai) Website: forecasting-workshop.github.io
    Image
    Image
    197K0197K
  • user avatar
    MiroMindAI
    @miromind_ai
    Apr 13
    🎉 @BinWang_Eng Great Hackathon event—what an excellent theme! Couldn't resist giving MiroThinker a try😁dr.miromind.ai/share/9746729d…
    user avatar
    Bin Wang
    @Bin_Wangg
    Apr 12
    Attended the UCWS Singapore Hackathon 2026 yesterday, the topic was on OpenClaw x Finance. The theme: "How to Make Money with AI Agents?" — a question every founder building in this space is thinking about right now.
    Image
    Image
    Image
    Image
    2K02K
Advertisement
Advertisement