Load dataset We'll fine-tune Gemma 4 12B to master chess. Given a set of previous move (one move missing) & the final result it has to predict the missing move. In order to do this we're using the ChessInstruct dataset from HuggingFace. Check this
CODE
-
LoRA configuration for efficient fine-tuning with Unsloth
By
–
Define LoRA config We'll use LoRA for efficient fine-tuning. To do this, we use Unsloth's PEFT and specify:
– The model
– LoRA low-rank (r)
– Layers for fine-tuning (finetune_language_layers, finetune_attention_modules) Check this code -
Loading Gemma 4 12B model with Unsloth tokenizer
By
–
Load the model We start by loading Gemma 4 12B and its tokenizer using Unsloth. Check this
-
Fine-tuning Gemma 4 12B to master chess on 8GB VRAM
By
–
Google released Gemma 4 12B, a multimodal model that runs text, images, and audio on 8GB VRAM!
— Akshay 🚀 (@akshay_pachaar) 7 juin 2026
We'll fine-tune it to master chess and predict the exact next move.
Tech stack:
– @UnslothAI for efficient fine-tuning.
– @huggingface transformers to run it locally.
Let's go! 🚀 pic.twitter.com/qPXjlYrShsGoogle released Gemma 4 12B, a multimodal model that runs text, images, and audio on 8GB VRAM! We'll fine-tune it to master chess and predict the exact next move. Tech stack:
– @UnslothAI for efficient fine-tuning.
– @huggingface transformers to run it locally. Let's go! -
Anthropic security guide: stop trusting AI agents
By
–
ANTHROPIC JUST PUBLISHED A 36-PAGE SECURITY GUIDE AND THE MESSAGE IS CLEAR: STOP TRUSTING YOUR AI AGENTS. If you work with Claude Code, MCP servers, or any automation tool, keep reading. The time you had to react
-
Limitations of CLAUDE.md Files for Context Control
By
–
Great writeup! One observation I've had lately is that one could spend hours crafting the perfect CLAUDE.md or Agents.md, but this only controls how the agent writes the code. It doesn't control how much context the backend dumps into the conversation on every tool call.
-
Reflection on Codex non-use: missing context, skills, forgetfulness, and large overhang
By
–
Whenever I don’t use codex for a task, I ask myself why and usually realize that there’s some missing context, I needed to write a skill, or I just didn’t think to use it. Rarely is it because the task is outside of the capabilities of the model. Overhang right now feels large.
-
Hope for Codex Desktop controlling other desktop instances
By
–
hope we aren’t far from codex desktop controlling other desktop instances
-
CEO Billionaire Proposes Proof-Gated Evolution Constitution for AI: GoalOS
By
–
RSI was missing a constitution. Not “AI improves itself.” Proof-gated evolution: Aim → Act → Prove → Evolve. No proof, no evolution.
No eval, no propagation.
No rollback, no release. AEP-001: GoalOS https://
github.com/MontrealAI/pro
of-gradient/blob/main/docs/standards/AEP-001/GoalOS_Proof-of-Evolution-Constitution_AEP001_v12.1_Institutional_Final.pdf
… #GoalAI #MontrealAI #QuebecAI -
AI will augment coders, not replace them
By
–
The more I analyze the relationship between humans and AI in coding and code generation, the more I wonder. You know what I think? I feel that AI is not going to replace coders, it's going to augment them. Bad coders will become better, and the…