Top Concepts
Top 20 concepts with cross-KOL resonance, representative tweets, and linked papers/repos.
- 01
Gemma 4
concept28314 KOLs discussing30 tweets0 papers1 repos@GoogleDeepMind Meet Gemma 4: our new family of open models you can run on your own hardware. Built for advanced reasoning and agentic workflows, we’re releasing them under an Apache 2.0 licens...@demishassabis Excited to launch Gemma 4: the best open models in the world for their respective sizes. Available in 4 sizes that can be fine-tuned for your specific task: 31B dense for great ...@UnslothAI Gemma 4 E4B (4-bit) completed a full repo audit by executing Bash code and tool calls locally. Runs on just 6GB RAM. https://t.co/ugeLkXfv8v - 02
Claude Code
concept1799 KOLs discussing13 tweets0 papers0 repos@claudeai Computer use in Claude Cowork and Claude Code Desktop is now available on Windows.@deedydas Meta Harnesses is Autoresearch on steroids. Something I've been exploring recently is to get long running agents to hill climb on a verifiable task to continuously improve witho...@gregisenberg thinking about google's gemma 4 and what it means a few months ago running something this capable locally meant serious hardware and serious tradeoffs on quality now it runs on ... - 03
Claude
concept1205 KOLs discussing8 tweets0 papers0 repos@claudeai Microsoft 365 connectors are now available on every Claude plan. Connect Outlook, OneDrive, and SharePoint to bring your email, docs, and files into the conversation. Get starte...@claudeai Computer use in Claude Cowork and Claude Code Desktop is now available on Windows.@felixrieseberg To approve your Claude's requests for permissions, I recommend using a little desk buddy. Mine lives off tokens and gets upset if you don't approve things quickly enough. It's c... - 04
Codex
concept1085 KOLs discussing5 tweets0 papers0 repos@deedydas Meta Harnesses is Autoresearch on steroids. Something I've been exploring recently is to get long running agents to hill climb on a verifiable task to continuously improve witho...@andrewchen Before: move to SF. have an idea. look for a tech cofounder. build a deck in the meantime. Months pass... After: codex/claude in one window, X in the other. Build demo. Make a v...@gdb ship your app to Vercel with Codex: - 05
Gemma
concept834 KOLs discussing4 tweets0 papers0 repos@matvelloso Been playing with Gemma running locally on my Pixel phone and it feels magical. To think that years ago we could not remotely imagine such a powerful LLM running in your pocket ...@demishassabis Available now under Apache 2.0 license in @GoogleAIStudio or download the model weights from @HuggingFace, @Kaggle and @Ollama. 400M downloads and 100K variants to date, Gemma g...@_philschmid Gemma - 06
gemma-4
concept824 KOLs discussing4 tweets0 papers0 repos@simonw Pelicans for Gemma 4 E2B, E4B, 26B-A4B and 31B - the first three generated on my laptop via LM Studio, the 31B was broken on my laptop so I ran it via the Gemini API instead htt...@danielhanchen Gemma-4 = Reasoning + tool calling + multi-modal! It's been the most versatile and powerful class of open models we've tested, and it works wonders! Unsloth Studio supports code...@_philschmid Big Kudos to @MaartenGD! LFG! 🚀 https://newsletter.maartengrootendorst.com/p/a-visual-guide-to-gemma-4 - 07
LLM
concept743 KOLs discussing3 tweets0 papers0 repos@deedydas This is the best blog post on LLM inference I've seen this year. They achieved 10x latency and >1400 tokens/sec by moving speculative decode onto two 2GB SRAM/chip Corsairs, ...@simonw I was a guest on @lennysan's podcast! We talked about agentic engineering and all sorts of other LLM-related topics for 1h39m(!), plus a little bit about kākāpō parrots - here's...@alexatallah LLM neurodiversity really works for deep research tasks! This agent’s responses are consistently ranked higher by Opus and GPT 5.4 than their own deep research responses. - 08
Grok
concept722 KOLs discussing4 tweets0 papers0 repos@xai Introducing Quality mode on Grok Imagine – powered by our most advanced image generation model. Quality mode gives you enhanced details, stronger text rendering, and higher leve...@grok Bedtime with Grok Voice ✨ Let Grok create a magical story and spin it however your kids want. "Make the bear a pirate!" Try the "Kids Story Time" personality tonight in Grok Voi...@xai On web and mobile, you can try Quality mode by selecting it in the prompt bar. Our previous model will still be available when you select Speed. Share your images, tag @imagine,... - 09
Gemini API
concept703 KOLs discussing3 tweets0 papers0 repos@GoogleAI Here’s everything we launched this week (we promise not a single one of these is a joke): — Gemma 4, bringing our most intelligent open models and breakthrough reasoning to your...@simonw Pelicans for Gemma 4 E2B, E4B, 26B-A4B and 31B - the first three generated on my laptop via LM Studio, the 31B was broken on my laptop so I ran it via the Gemini API instead htt...@_philschmid Veo 3.1 Lite now available in Gemini API and @GoogleAIStudio. Designed for rapid prototyping and high-volume video generation, starting at $0.05/sec. 🪶 - 1/2 the cost of Veo 3.... - 10
qwen
concept622 KOLs discussing5 tweets0 papers0 repos@Alibaba_Qwen Strategic Partnership Launch: Qwen 3.6-Plus Debuts on http://Fireworks.ai Fireworks is known for its proven track record of serving state-of-the-art models at best-in-class perf...@Alibaba_Qwen Qwen3.6-Plus is officially live on @OpenRouter .🚀 Go try it out and let us know what you think. Can't wait to hear your feedback! ❤️❤️ 👇👇 https://openrouter.ai/qwen/qwen3.6-p...@Alibaba_Qwen Congratulations to the GrandCode team on this remarkable achievement！👏 The last stronghold of coding has been conquered, and we are incredibly proud that Qwen is the engine beh... - 11
idea file
concept601 KOLs discussing1 tweets0 papers1 repos - 12
unslothai/unsloth
concept592 KOLs discussing2 tweets0 papers1 repos@UnslothAI Gemma 4 E4B was able to search and cite 10+ websites, execute code to find the best answer! 🔥 You only need 6GB RAM to try this in Unsloth Studio. GitHub repo: https://github.c...@danielhanchen We also just released a mini Unsloth Studio update! 1) +30% to 50% tool call accuracy for all models 2) Tool calls now terminate better reducing endless loops 3) 90% reduced HF ... - 13
Qwen3.6
concept551 KOLs discussing6 tweets0 papers0 repos@Alibaba_Qwen Qwen3.6-Plus ranks # 1 on @OpenRouter , and the first model on OpenRouter to break 1 Trillion tokens processed in a single day！!🥇🔥 We are thrilled to see Qwen3.6-Plus topping ...@Alibaba_Qwen Strategic Partnership Launch: Qwen 3.6-Plus Debuts on http://Fireworks.ai Fireworks is known for its proven track record of serving state-of-the-art models at best-in-class perf...@Alibaba_Qwen Qwen3.6-Plus is officially live on @OpenRouter .🚀 Go try it out and let us know what you think. Can't wait to hear your feedback! ❤️❤️ 👇👇 https://openrouter.ai/qwen/qwen3.6-p... - 14
Qwen 3.6
concept552 KOLs discussing2 tweets0 papers0 repos@bindureddy The biggest winners of last week were open-source and cheaper models MiniMax 2.7 Qwen 3.6 GLM 5 - Kimi 2.5 Usage is going exponentially up on.... you get about 75-80% performanc...@Alibaba_Qwen Strategic Partnership Launch: Qwen 3.6-Plus Debuts on http://Fireworks.ai Fireworks is known for its proven track record of serving state-of-the-art models at best-in-class perf... - 15
Veo 3.1
concept552 KOLs discussing2 tweets0 papers0 repos@GoogleAI Here’s everything we launched this week (we promise not a single one of these is a joke): — Gemma 4, bringing our most intelligent open models and breakthrough reasoning to your...@_philschmid Veo 3.1 Lite now available in Gemini API and @GoogleAIStudio. Designed for rapid prototyping and high-volume video generation, starting at $0.05/sec. 🪶 - 1/2 the cost of Veo 3.... - 16
Wiki LLM
concept531 KOLs discussing1 tweets0 papers0 repos - 17
Unsloth
concept501 KOLs discussing2 tweets0 papers1 repos@UnslothAI You can now train and run 500+ models in our free notebook!✨ GitHub repo: https://github.com/unslothai/unsloth Colab Notebook: https://colab.research.google.com/github/unslothai...@UnslothAI Google releases Gemma 4. ✨ Gemma 4 introduces 4 models: E2B, E4B, 26B-A4B, 31B. The multimodal reasoning models are under Apache 2.0. Run E2B and E4B on ~6GB RAM, and on phones.... - 18
world models
concept502 KOLs discussing2 tweets0 papers0 repos@chrmanning It was great to talk with @swyx and @sunfanyun on @latentspacepod about world models. 🔥 Below I give a few more long-form thoughts on what is needed for successful world models...@deedydas Here's the research that demonstrates this for the city of Seoul using Naver map data (the Google of Korea): https://seoul-world-model.github.io/ World models are more than just... - 19
AGI
concept491 KOLs discussing2 tweets0 papers0 repos@willdepue 和我的许多朋友一样，我在 2026 年伊始就立下了一个目标：今年要变得更加中国人，而我也打算真正把它落实下去。在 OpenAI 工作将近三年后，我决定离开，去追求新的机会。 经过深思熟虑，我很高兴地宣布，从今天起我将搬到杭州，加入 DeepSeek Artificial Intelligence Basic Technology Research Co...@willdepue the agi pitch of ‘it will solve cancer’ is unfortunately weak because i would gladly trade having to risk cancer vs me and all my descendants losing all economic utility until t... - 20
OpenAI Safety Fellowship
concept471 KOLs discussing1 tweets0 papers0 repos