211 summarised stories about Open Source Models, each linking back to the original source. Browse all topics →
TechCrunch 1 week ago 46 ● 3 sources
Nous Research raised a $90 million Series B to confirm its $1.5 billion valuation and launched “Hermes for Businesses.” The funding round values the company at $1.5 billion. It will push Nous into enterprise deployments of customized multi-step AI agents that keep customer data private and secure, aiming to grow revenue beyond $100 million in 2026.
Product Hunt 1 week ago 30
Hallmonitor is launching as a free, open-source app that shows Claude Code and Codex coding-agent sessions on a Mac and servers, including in the menu bar and terminal. It has “7 followers.” As a result, users can monitor which tasks are done or stuck and answer questions from the notch.
Google 1 week ago 1 ● 7 sources
Google launched EmbeddingGemma 2, a private multimodal on-device search embedding model that maps text, images, audio, and video into one embedding space. It has 740 million parameters and supports an 8K token context window. Developers can now build fully offline cross-modal semantic search and retrieval (including on-device RAG) using the new open Apache 2.0 model and its smaller, truncated vector outputs for local storage.
Product Hunt 1 week ago 17
OpenCharm launched as an open-source hardware “body” that sits by a Mac’s notch to control an AI agent hands-free via a single key press. The listing says you “Hold one key” to talk so tools like Claude Code or Codex can start and ask for input. It changes how you interact with already-running AI agents by turning voice and handoff prompts into a one-key physical workflow.
Allen Institute (AI2) 1 week ago 8
Ai2 published the Bolmo research in Nature and released new Hugging Face checkpoints for byte-level language models beyond the original Olmo-based versions. The release includes Stage 1 checkpoints that freeze the global model while training Bolmo’s byte-level components. This enables faster retrofitting of subword models into competitive byte-level models and supports broader experimentation without retraining from scratch.
MarkTechPost 1 week ago 16
Meta open-sourced Rebalancer, a C++ (with Python interface) library for solving constrained assignment problems, and provides installation packages plus a debugging UI. The release is available via PyPI as version 1.0.4 for Python 3.12+ and reports a P99 solve time of 12 seconds on 265k objects and 3.2k bins at Meta. Rebalancer changes resource-allocation workflows by letting teams specify problems separately from the solving method, then using local search or MIP backends depending on problem size.
MarkTechPost 1 week ago 45
Laya, Convai Innovations’ open-source non-autoregressive decision engine, is evaluated in a tutorial by running typed multiple-choice/score/yes-no queries and checking zero-shot behavior on labelled CLINC150 banking data. The guide specifies a 421-million-parameter encoder that scores all options in a single forward pass with zero output tokens. It reports how option wording/order, probability calibration via temperature, and an abstention gate affect accuracy and the trustworthiness of the returned probabilities for a production-style router.
The daily briefing
TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.
TLDRocket
18 October 2026
AI oversight stopped being a PowerPoint exercise today and got hands-on. Anthropic said it has given outside AI watchdog METR “employee-like” access, embedding a METR team in its offices with access to hardware and internal systems so risk evaluation can run closer to the actual model and infrastructure—not just post-hoc reports. That trend of tightening control shows up elsewhere: U.S. banks and brokerages are blocking outside AI agents from logging in and moving money, with Charles Schwab reporting it blocks agents “in the tens and twenties each week,” while routing limited balance viewing and analysis through Schwab’s own MCP servers. Even the courts are joining in—at least eight states sued OpenAI, alleging ChatGPT helps “cyberhacking agents” and misleads adults, raising the stakes for how tightly companies must constrain agent behavior.
Read the full briefing →