1. Unsloth AI • Fine-tune models like Qwen3, Llama 4, and Gemma 3 up to 2× faster with 70% less VRAM
• Supports low-resource setups and runs on consumer GPUs or even Colab/Kaggle with ~3 GB VRAM GitHub repo →
LLMS
-
Unsloth AI: Fine-tune LLMs 2× faster with 70% less VRAM
By
–
-
Four Open-Source Libraries to Accelerate LLM Fine-Tuning
By
–
4 libraries that accelerates fine-tuning of LLMs (100% open-source)
-
Qwen3-235B-A22B-Thinking-2507 model released
By
–
Qwen released a new Qwen3-235B-A22B-Thinking-2507 model, which is now available on Qwen Chat as well!
-
Comparison of AI-powered development tools for builders
By
–
Most devs are chasing GitHub Spark. Builders are using this.
-
New AI Models Compared: Lobster, Nectarine, Starfish Performance
By
–
New models on the @lmarena_ai WebDev arena:
— Peter Gostev (@petergostev) 25 juillet 2025
– Lobster
– Nectarine
– Starfish (not in this video)
In the video compared to the 'Anonymous Chatbot' (aka o3-Alpha) from 17th July.
Observations:
– Lobster is closest to the o3-Alpha, but nowhere near as good
– Nectarine was… pic.twitter.com/BmSaoUXmuANew models on the @lmarena_ai WebDev arena: – Lobster – Nectarine – Starfish (not in this video) In the video compared to the 'Anonymous Chatbot' (aka o3-Alpha) from 17th July. Observations: – Lobster is closest to the o3-Alpha, but nowhere near as good – Nectarine was
-
Technical Walkthrough for AI Model Fine-Tuning
By
–
Fine tuning walkthrough alone makes it worth bookmarking.
-
GPT-5 Launch, AI Protein Design, Productivity Tools, Jobs
By
–
Top stories in AI today: – OpenAI readies GPT-5 for August debut
– AI designs cancer-killing proteins in weeks
– Turn Comet into an AI-powered productivity assistant
– Microsoft maps how workers use AI
– 4 new AI tools & 4 job opportunities Read more: https://
therundown.ai/p/openai-readi
es-gpt-5-for-august-debut
… -
BERT CPU Model Powers Smart Home Interface with Function Calling
By
–
It's super fun to use a BERT-sized model that can run on a CPU and power a smart home interface. Also, we ranted a bit about function calling templates with a comparison between JSON and Python. Hope you enjoy it! 🙂
-
Using Claude as a 24/7 senior engineer assistant
By
–
Using Claude like this feels like having a senior engineer on-call 24/7. Way beyond what Copilot offers.
-
LFM2 Constrained Generation Smart Home Control Demo
By
–
LFM2 Outlines Constrained generation is amazing to squeeze more performance from tiny models. We just released a blog post with @remilouf with a smart home control demo. We used LFM2-350M to save notes, close blinds, and handle displays.