🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

introducing: picoDeepResearch multi-turn tool use + soft rewards + self-play + GRPO You define the arena (report prompts + judging principles) the model generates reports, uses tools (web search), then competes in round-robin battles judged by an LLM winner gets the gradient

👌 It’s always cool to see some Weaviate code popping up during Google I/O 👉 youtu.be/kj80m-umOxs

I published Vol. 105 of "Top Information Retrieval Papers of the Week" on Substack. 🔗 recsys.substack.com/p/training-mul…

[LG] DataRater: Meta-Learned Dataset Curation D A. Calian, G Farquhar, I Kemaev, L M. Zintgraf... [Google DeepMind] (2025) arxiv.org/abs/2505.17895

There are really only 𝟮 𝘄𝗮𝘆𝘀 to level up your AI agent's performance. You can either: 1️⃣ Choose a better model 2️⃣ Optimize your prompts That's it. Everything else is just details. For model selection, I've tested dozens of options, and 𝗚𝗲𝗺𝗶𝗻𝗶 consistently stands out

Thanks to Omar Khattab's and Cyrus's suggestions we now have a special module for dspy within attachments. No need for .dspy() everywhere😍 Also, don't be fool by the simplicity of this code. This code is a fully working llm program! This program loads power points

You don’t need frontier lab resources for frontier lab automated LLM evaluation. To prove this, we’re open-sourcing j1-nano and j1-micro: two absurdly tiny (600M & 1.7B parameters) but mighty reward models competitive with orders-of-magnitude larger peers. j1-nano and j1-micro

Multi-Vector Retrieval methods such as ColBERT / ColPali are super powerful *but* often too costly to store and search! Well... not anymore! Introducing MUVERA! 🚀 I am SUPER EXCITED to publish our newest Weaviate Podcast on MUVERA with Rajesh Jayaram (Rajesh Jayaram) from

YouTube: youtu.be/nSW5g1H4zoU Spotify: creators.spotify.com/pod/show/weavi…

Weaviate Podcast #123 is live! MUVERA! 🚀

If you're in SF in 2 weeks, see you at the @Databricks Data + AI Summit, where we have a long list of sessions on DSPy, including: ■ DSPy 3.0 by Krista Opsahl-Ong & me (Jun 11) ■ DSPy <> MLflow by Chen Qian (Jun 11) ■ Opinionated intro to DSPy by Drew Breunig (Jun 10)

We've just released Rank-R1-v0.2, based on Qwen3-32B. It's now the SOTA on the BRIGHT leaderboard!

Apache Spark 4.0 is out with some huge improvements across the board. SQL’s much more powerful, Spark Connect makes it easier to run apps, new languages and more. It’s amazing to see the community still growing fast and releasing over 5000 patches in 4.0. databricks.com/blog/introduci…

frontier ai today

#DennysPick 🌟 If you’re building GenAI applications with DSPy and want full observability, seamless debugging, and true MLOps best practices, this session is a must-see at #DataAISummit! I’m excited to recommend “Streamlining DSPy Development: Track, Debug, and Deploy With

When people say “multi-agent”, 80% of the time they just want to write a *multi-module* DSPy program, in which case it’s been supported since 2022! 15% of the time they want to give dspy.ReAct `tools` that are themselves agents, in which case it’s been supported since Oct 2023.

bro come on