AI & Tech Developments - Mar 28

breaking news big tuna ai labs presents Qwen3.5-9b-Sushi-Coder sft on opus reasoning traces and codeforces with unsloth then RLed with hermes agent using Atropos for coding RL im not raising !

— @tunahorse21

Pi prompt templates can now loop and rotate between different models on each iteration. In this example it runs each loop with a different model as a subagent based on a fork of the current chat. pi install npm:pi-prompt-template-model https://github.com/nicobailon/pi-prompt-template-model…

— @nicopreme

Build voice agents that do real work. We built a clinic concierge demo for a Singapore health clinic with gpt-realtime-1.5. It speaks naturally with patients, collects the right details, and books appointments in real time.

— @OpenAIDevs

GLM-5.1 by @Zai_org is one of the best agentic models out there. I've been testing it early and it is genuienly impressive. Way better at instruction following, long running tasks than previous generation. full review here: https://youtu.be/UxGieu7PaPg

— @aicodeking

We have integrated @huggingface as a first-class inference provider in Hermes Agent. When you select Hugging Face in the model picker it now shows 28 curated models organized by use case, with a custom option for the 100+ other models they serve.

— @NousResearch

For everyone with 16 or less GB ram that want to run a coding model locally. This is an autocomplete code model that I’ve noticed does pretty well. Not as good as Cursor tab but good enough

— @0xSero

GIVE ME YOUR ATTENTION FOR LIKE 5 SECONDS GITHUB IS GONNA START USING *YOUR* CODE AND DATA TO TRAIN AI USING COPILOT!!!!!!!!!!! DISABLE IT AT http://github.com/settings/copilot/features…

— @_kaivara_

Qwen 3.5 27B (Dense) with Hermes Agent is REALLY GOOD

— @TheAhmadOsman

Introduction of Qwen 3.5-9b-Sushi-Coder by Big Tuna AI Labs.

— @tunahorse21

Pi prompt templates can now loop and rotate between different models.

— @nicopreme

Development of a clinic concierge demo using GPT technology.

— @OpenAIDevs

Discussion on the capabilities of GLM-5.1 model.

— @aicodeking

Announcement of improvements in Hermes Agent integration with Hugging Face.

— @NousResearch

Overview of a coding model suitable for low RAM devices.

— @0xSero

Warning about GitHub's use of user code and data for AI training.

— @_kaivara_

Positive feedback on Qwen 3.5 27B model performance.

— @TheAhmadOsman

TurboQuant CUDA for llama.cpp: 3.5x KV cache compression that beats q8_0 quality.

— @spiritbuun

CLI agents using CLI browsers, potentially the future of AI browser use.

— @shawn_pana