What if every desktop app could rewrite itself, locally, for free?
Built Loopa — a small SDK where you type a feature request inside a running app, and a local LLM regenerates the next version. Validates, hot-swaps, rolls back on crash. No cloud, no API bill.
deepseek v4 - vllm . (APIServer pid=23512) Value error, The checkpoint you are trying to load has model type `deepseek_v4` but Transformers does not recognize this architecture. #deepseek#deepseekv4. nightly builds
Experimenting with Apple’s Foundation Model from the command line?
I made fm, an interactive macOS terminal chat tool for the built-in model.
brew tap kosminus/fm && brew install fm
https://t.co/eIEW6LKpcm
ai layer over linux / ubuntu . an event-driven runtime that sits between your operating system and LLM agents via Ollama. similar with #openclaw#linux#ollama#ubuntu#llm#openclaw https://t.co/Eqiju3SAnm
AI Event Bus. A local-first intelligence bus for orchestrating multiple LLM agents via Ollama. Event-driven architecture where producers emit events, routing rules match them to LLM consumers, and agent outputs flow back as new events. https://t.co/TUl0PsCwqS
Romanian AI in real operations.
Nemesis AI, developed by the Oves Enterprise team, is now integrated into the EAGLS™️ counter-drone system used by the U.S. military in the Middle East.
Proud to build technology that helps detect, track and intercept aerial threats in real time.
Just open-sourced ReasonFlow — build LLM pipelines as Python DAGs with >> chaining and parallel() branches. 4 node types, auto cost tracking, budget enforcement, and 13 ready-to-run examples. Works with OpenAI, Anthropic, Google, and Ollama. https://t.co/cBSSs8yply
I built QueryWise — an open-source text-to-SQL app. Ask questions in plain English, get SQL + results.
Works with Claude, OpenAI, or Ollama (fully local). Supports PostgreSQL, BigQuery, Databricks.
https://t.co/IS6y7e6U0T
https://t.co/DrCb3xMuy9
#opensource#llm#sql#ai