AI
165 items on noze.it
Products
ProductAdmina Enterprise
The open Admina platform + enterprise services for governed, sovereign AI: supports NIS2 and AI Act compliance.
ProductAgenticHealth
On-premise agentic clinical AI. MDR pathway planned, currently Research Use Only.
ProductDataGovern
The agentic AI that oversees your documents and processes and prepares NIS2, GDPR and AI Act compliance.
ProductIntelliPA
Turnkey self-hosted AI for Italian municipalities and public bodies.
ProductVerdePilot
The AI copilot for environmental and energy compliance at sites in Italy.
R&D Projects
ALA — Advanced Laboratory Automation
Joint R&D INPECO + BioRobotics Sant'Anna Pisa (EUR 3.5M, 30 researchers): automation and robotics for clinical labs, dee...
Dake — Redefining the Physics of Energy
Deep-tech research at the intersection of quantum physics, AI and energy engineering. AI-driven optimisation of energy g...
DebugABot — Debugging Autonomous Intelligence
Deep-tech infrastructure for governing autonomous AI agents and embodied intelligent systems. Kill switch, model fingerp...
MindRACES — anticipatory cognitive systems
EU FP6 project on anticipatory cognitive mechanisms: building systems able to predict the outcome of their own actions....
Articles (148)
16 July 2026Codex Micro and the AI hardware that works: augment, don't replace
OpenAI launches the Codex Micro, a $230 macropad (with Work Louder) to drive fleets of Codex agents: its first branded h...
15 July 2026Inkling: Thinking Machines' open-weights model, built for fine-tuning
Thinking Machines, Mira Murati's lab, releases Inkling: its first model, a 975-billion-parameter Mixture-of-Experts (41 ...
14 July 2026What your AI coding CLI actually sends: the Grok case and the paradox in reverse
A wire-level analysis of xAI's coding CLI (grok) documents the upload of the entire repository and .env files to an xAI ...
14 July 2026colibri: GLM 5.2 (744B) on 25 GB of RAM, streaming the experts from disk
colibri, a C inference engine released by JustVugg (Apache 2.0), runs GLM 5.2, a 744-billion-parameter MoE model, on a c...
12 July 2026Ornith 1.0: open models that learn their own scaffold for agentic coding
DeepReinforce has released Ornith 1.0, an open-weight family of models for agentic coding, from 9B Dense up to 397B MoE,...
10 July 2026Agentic loop (3): security and governance
To work, the loop reads untrusted content into the context and has tools to act: that is its attack surface. Part 3 of 3...
9 July 2026Harness engineering: the environment that makes agents reliable
The model decides what to write; the harness governs when, where and how. Harness engineering is the discipline of build...
8 July 2026Agentic loop (2): context, patterns and multi-agent
The basic cycle is simple; making it last is not. Part 2 of 3: context engineering (truncation, compaction, externalized...
7 July 2026Claude's J-space: reading the model's unspoken thoughts
On 6 July 2026 Anthropic published 'A global workspace in language models': inside Claude there is a small internal work...
6 July 2026Agentic loop (1): anatomy of the cycle
An LLM is stateless: an agent is what you get when you put it in a loop with a goal, tools, observations and a stop cond...
2 July 2026Fable 5 redeployed: what changes for cybersecurity
On 1 July 2026 Anthropic restored access to Claude Fable 5 (and Mythos 5 for US defenders only, via Project Glasswing) a...
30 June 2026DwarfStar 4 by antirez: quasi-frontier AI, local and open
antirez (Salvatore Sanfilippo, creator of Redis) has released DwarfStar 4 (ds4): a native, self-contained, open source (...