AI News
2.3K views | +35 today
 
Scooped by Philippe Gassmann
onto AI News
September 18, 3:02 AM
Scoop.it!

Weave Engineering Intelligence: ML Models to Measure & Optimize Engineering Work

Weave Engineering Intelligence: ML Models to Measure & Optimize Engineering Work | AI News | Scoop.it
Weave understands engineering work by combining LLMs and domain-specific machine learning. We tell you how much work is getting done, how good it is and how to optimize your token allocation. Used by startups and the fortune 100 (YC W25).
No comment yet.
Your new post is loading...
Your new post is loading...
Scooped by Philippe Gassmann
Today, 4:02 AM
Scoop.it!

Exclusive: Italian robotics startup Mosaico raises €7.5m to tackle AI’s data problem

Exclusive: Italian robotics startup Mosaico raises €7.5m to tackle AI’s data problem | AI News | Scoop.it
Italian startup aims to make it easier for robotics companies to organise data needed to train AI models.
No comment yet.
Scooped by Philippe Gassmann
Today, 3:02 AM
Scoop.it!

Intergrid lands €23.5M to develop a ‘self-driving car’ for factory energy management — TFN

Intergrid lands €23.5M to develop a ‘self-driving car’ for factory energy management — TFN | AI News | Scoop.it
Intergrid secures €23.5M in total funding to scale Autopilot, its AI-driven platform for autonomous industrial energy management across more than 100 sites.
No comment yet.
Scooped by Philippe Gassmann
Today, 3:02 AM
Scoop.it!

Agora Labs: 4,3 milioni per l'AI Made in Italy nella sanità

Agora Labs: 4,3 milioni per l'AI Made in Italy nella sanità | AI News | Scoop.it
Agora Labs raccoglie 4,3 milioni per una piattaforma di AI federata che rende i dati clinici utilizzabili nella ricerca senza trasferirli dagli ospedali.Frasi secondarie: intelligenza artificiale federata, dati clinici, HealthTech italiana
No comment yet.
Scooped by Philippe Gassmann
Today, 3:02 AM
Scoop.it!

s3onghyun/agent-otel-kit: One-command OpenTelemetry observability kit for AI coding agents (Claude Code, Codex, Gemini CLI): collector, Prometheus/Loki/Tempo, Grafana and Perses dashboards ·

s3onghyun/agent-otel-kit: One-command OpenTelemetry observability kit for AI coding agents (Claude Code, Codex, Gemini CLI): collector, Prometheus/Loki/Tempo, Grafana and Perses dashboards · | AI News | Scoop.it
One-command OpenTelemetry observability kit for AI coding agents (Claude Code, Codex, Gemini CLI): collector, Prometheus/Loki/Tempo, Grafana and Perses dashboards - s3onghyun/agent-otel-kit
No comment yet.
Scooped by Philippe Gassmann
Today, 3:02 AM
Scoop.it!

Attest

Attest | AI News | Scoop.it
Test what your AI agent says, which tools it calls and what it changes. Replay recorded model responses and enforce checks in CI.
No comment yet.
Scooped by Philippe Gassmann
Today, 3:02 AM
Scoop.it!

Chema Garabito on X: "@levelsio you can use https://t.co/e0SgU8dWmg from @speridlabs, it is apache 2.0, give photorealistic quality
and it is under 1b parameters, runs in under 8gb of vram and it ...

@levelsio you can use https://t.co/e0SgU8dWmg from @speridlabs, it is apache 2.0, give photorealistic quality

and it is under 1b parameters, runs in under 8gb of vram and it is SOTA + threejs has native support for this high quality gaussian splats :)
No comment yet.
Scooped by Philippe Gassmann
Today, 3:01 AM
Scoop.it!

Emrah Hizarci on X: "qwen4-max is out. alibaba says tasks cost like 60% less than with their last flagship, their own numbers though
the part I care about: smaller open-weight qwen4 models are com...

qwen4-max is out. alibaba says tasks cost like 60% less than with their last flagship, their own numbers though

the part I care about: smaller open-weight qwen4 models are coming, including a 27B you can run locally. that's the one I'm waiting for
No comment yet.
Scooped by Philippe Gassmann
Today, 3:01 AM
Scoop.it!

[2611.18471] Do Verifier Judgments Survive Abstraction? Measuring Verification Transport for LLM Agent Evaluation

[2611.18471] Do Verifier Judgments Survive Abstraction? Measuring Verification Transport for LLM Agent Evaluation | AI News | Scoop.it
LLM-based verifiers are increasingly used to evaluate agent executions, but their reliability is typically measured under a single representation of the trajectory. In deployed agent systems, the same execution may be exposed through composed tools, delegation boundaries, or compacted histories, changing what execution evidence is directly observable to the verifier. It therefore remains unclear whether correct verifier judgments persist across such abstraction boundaries. We introduce Verification Transport, a reliability test that measures judgment stability under abstraction-induced changes in execution observability. We construct matched native, abstracted, and restored views of the same frozen executions while holding the task, realized trajectory, outcome, ground truth, evaluation criterion, and verifier fixed. Across four agent-evaluation benchmarks, seven verifier backbones, four verification protocols, and three abstraction boundaries, we find substantial transport failure, reaching about 31% under macro abstraction and delegation. Verification protocols exhibit distinct failure modes: direct verification more often produces incorrect verdicts, whereas evidence-grounded and deliberative protocols tend to abstain under limited observability. This suggests that lower transport failure rates do not necessarily imply greater reliability. Restoring the corresponding execution provenance recovers 85.5% of transport failures, while recovery is lower when provenance is lossy or unavailable. These results show that native-view accuracy alone is insufficient to characterize verifier reliability when execution evidence crosses abstraction boundaries.
No comment yet.
Scooped by Philippe Gassmann
Today, 3:01 AM
Scoop.it!

[2611.18393] WeatherDreamer: Learning a World Model from Multimodal Earth Observations

[2611.18393] WeatherDreamer: Learning a World Model from Multimodal Earth Observations | AI News | Scoop.it
Machine learning weather models now rival operational global forecasts but still start from analyses produced by separate, expensive data assimilation (DA) systems, which for global models are typically updated every six hours. Observation-to-observation models avoid this dependence, but most omit the recursive filtering cycle that lets DA systems carry a background state through sparse, irregular observations. We introduce WeatherDreamer, a world model that learns this cycle end-to-end from 18 global observing systems, without reanalysis. It infers a latent state from observations, carries it forward hour by hour with learned dynamics, corrects it each hour with new observations, and can predict what any of its observing systems would measure at any location. At held-out stations, its ensembles beat NOAA's Global Ensemble Forecast System (GEFS) for temperature, dew point and wind out to 12 h and are better calibrated than GEFS and the ECMWF ensemble. Precipitation, cloud-top height and solar irradiance, never inputs at inference, are inferred with skill matching or exceeding operational products at short leads. Against the first guess of NOAA's Gridpoint Statistical Interpolation (GSI) DA system, WeatherDreamer's short-range departures are within 10% of GSI's at surface stations and within 25% for moisture-sensitive microwave channels, but up to five times larger for temperature-sounding channels. Never given observation errors, its learned update assumes errors close to known instrument and retrieval errors of microwave sounders and scatterometer winds. WeatherDreamer also separates viewing geometry from atmospheric state, recovering 83-96% of observed limb darkening and producing seamless nadir-view imagery. Producing an hourly global analysis in about 1.5 minutes on one GPU, WeatherDreamer shows that assimilation and forecasting can be learned together directly from observations.
No comment yet.
Scooped by Philippe Gassmann
Today, 3:01 AM
Scoop.it!

Indian chip startup Bharat Semi lands $24M seed round as it eyes global AI sensor market

Indian chip startup Bharat Semi lands $24M seed round as it eyes global AI sensor market | AI News | Scoop.it
Bharat Semi says it has secured design wins for its AI-enabled sensor technology with companies including General Atomics.
No comment yet.
Scooped by Philippe Gassmann
November 15, 2:02 AM
Scoop.it!

Moda, a Milano debutta Shopping Cube: il camerino virtuale guidato dall'intelligenza artificiale - Moda

Moda, a Milano debutta Shopping Cube: il camerino virtuale guidato dall'intelligenza artificiale - Moda | AI News | Scoop.it
Anteprima mondiale al Forum Retail 2026 per il cubo digitale a led ideato da Ottavia Pittaluga: si provano i vestiti senza cambiarsi e si acquista online o in negozio
No comment yet.
Scooped by Philippe Gassmann
November 15, 2:02 AM
Scoop.it!

ai-timon/timon ·

ai-timon/timon · | AI News | Scoop.it
Contribute to ai-timon/timon development by creating an account on GitHub.
No comment yet.
Scooped by Philippe Gassmann
November 15, 2:02 AM
Scoop.it!

Vaibhaav-Tiwari/deepseek-v3-jax: DeepSeek V3 (MLA, DeepSeekMoE, aux-loss-free balancing, MTP) in 190 lines of pure JAX, trained on a MacBook Air. Write-up: vaibhaav.in/writing/deepseek-v3-jax ·

Vaibhaav-Tiwari/deepseek-v3-jax: DeepSeek V3 (MLA, DeepSeekMoE, aux-loss-free balancing, MTP) in 190 lines of pure JAX, trained on a MacBook Air. Write-up: vaibhaav.in/writing/deepseek-v3-jax · | AI News | Scoop.it
DeepSeek V3 (MLA, DeepSeekMoE, aux-loss-free balancing, MTP) in 190 lines of pure JAX, trained on a MacBook Air. Write-up: vaibhaav.in/writing/deepseek-v3-jax - Vaibhaav-Tiwari/deepseek-v3-jax
No comment yet.
Scooped by Philippe Gassmann
Today, 3:02 AM
Scoop.it!

Germany’s Daisytuner raises €6 million to make AI models run faster across NVIDIA, AMD and emerging chips

Germany’s Daisytuner raises €6 million to make AI models run faster across NVIDIA, AMD and emerging chips | AI News | Scoop.it
Daisytuner, a Darmstadt-based DeepTech startup working on automating AI model performance engineering across any chip, has raised €6 million in Seed
No comment yet.
Scooped by Philippe Gassmann
Today, 3:02 AM
Scoop.it!

Il bilancio degli Stati Generali della Scuola Digitale 2026

Il bilancio degli Stati Generali della Scuola Digitale 2026 | AI News | Scoop.it
Oltre 5.000 presenze a Varese per la decima edizione della manifestazione promossa da Impara Digitale. Tra la nuova alleanza scientifica con il CNR, la sperime…
No comment yet.
Scooped by Philippe Gassmann
Today, 3:02 AM
Scoop.it!

Tthanh2k6/turbo-browser-mcp: Playwright MCP server built to cut agent turns and tokens: one-call form filling, delta snapshots, batched steps, and a submit guard on by default ·

Tthanh2k6/turbo-browser-mcp: Playwright MCP server built to cut agent turns and tokens: one-call form filling, delta snapshots, batched steps, and a submit guard on by default · | AI News | Scoop.it
Playwright MCP server built to cut agent turns and tokens: one-call form filling, delta snapshots, batched steps, and a submit guard on by default - Tthanh2k6/turbo-browser-mcp
No comment yet.
Scooped by Philippe Gassmann
Today, 3:02 AM
Scoop.it!

gavel: guided code review with your coding agent

gavel: guided code review with your coding agent | AI News | Scoop.it
Your coding agent prepares a brief and walks you through its changes. Question its choices beside the code, request revisions, and approve each review.
No comment yet.
Scooped by Philippe Gassmann
Today, 3:02 AM
Scoop.it!

db-branch/db-branch: Instant MySQL database branches on ZFS clones, clustered across VMs with ProxySQL routing ·

db-branch/db-branch: Instant MySQL database branches on ZFS clones, clustered across VMs with ProxySQL routing · | AI News | Scoop.it
Instant MySQL database branches on ZFS clones, clustered across VMs with ProxySQL routing - db-branch/db-branch
No comment yet.
Scooped by Philippe Gassmann
Today, 3:02 AM
Scoop.it!

Theo Oliveira on X: "Modelo aberto na Together (Ripple 65B): US$ 0,20 o milhão de token de entrada. GPT-6 Cyber da OpenAI: US$ 20. Cem vezes, man....
E tem SaaS de IA mandando até "resume esse text...

Modelo aberto na Together (Ripple 65B): US$ 0,20 o milhão de token de entrada. GPT-6 Cyber da OpenAI: US$ 20. Cem vezes, man....
E tem SaaS de IA mandando até "resume esse texto" pro modelo mais caro. Parabéns, a margem tá pagando a preguiça de rotear kkk
https://t.co/8EER8z3cqH
No comment yet.
Scooped by Philippe Gassmann
Today, 3:01 AM
Scoop.it!

[2611.18514] Can Mechanistic Interpretability Help Explain LLM Agent Failures? A Case Study on Web Element Hallucination

[2611.18514] Can Mechanistic Interpretability Help Explain LLM Agent Failures? A Case Study on Web Element Hallucination | AI News | Scoop.it
Mechanistic Interpretability (MI) has made significant progress in explaining the internal computations of Large Language Models (LLMs); however, these advances have yet to demonstrate comparable success in explaining the decision-making processes of LLM-based agents. This paper examines the challenges MI tools face when applied to realistic agentic failures. We focus on one such failure: web agents hallucinating actions that are not available on a website. We apply a series of mechanistic analyzes to Qwen2.5-7B-Instruct as it interacts with live websites. We find that a linear probe trained on the last prompt token can predict hallucinations before generation begins (0.80 AUC) and steering along a probe direction corrects 85\% of hallucinations, but disrupts 7\% of valid actions. However, these results only provide a partial explanation of why the agent hallucinates. The probe direction is sensitive to prompt formatting, losing much of its steering effect after a minor change to how actions are represented. Additionally, circuit identification can be difficult as the computation is distributed across multiple components and positions, and the information the agent relies on varies across examples. We discuss how these observations challenge the assumptions underlying standard MI methods for circuit discovery and feature identification, and outline future directions for making these methods better suited to explain failures in LLM-based agents.
No comment yet.
Scooped by Philippe Gassmann
Today, 3:01 AM
Scoop.it!

[2611.18553] WeightBench: Evaluating Language Models as Weight-Space Programmers

[2611.18553] WeightBench: Evaluating Language Models as Weight-Space Programmers | AI News | Scoop.it
Recent work shows that large language models (LLMs) can operate directly on the parameters of other neural networks. They can read what a network was trained to do, fix a network that fails, and write working networks from scratch, all without gradient descent. This capability has been termed Weight-Space Programming (WSP). Exploring this new paradigm systematically requires a robust benchmark. We introduce WEIGHTBENCH, a benchmark that evaluates LLMs of increasing capability on 42 WSP tasks across five tracks: reading, fixing and writing weights, recognizing edits that are impossible, and data-free compression. The networks whose weights the LLMs read and edit range from 2K-parameter control policies to 7B-parameter language models. Experimental benchmark results show that frontier agents can already read small networks and write working control policies without ever running them, but cannot yet patch language models, compress them without data, or tell an impossible edit from a hard one. On all 42 tasks, Claude Opus 5.5 reaches a macro score of 0.535 (Sonnet 5.5 0.466, Haiku 5.5 0.402, DeepSeek v4-flash 0.215); on the 38 tasks it was run on, GPT-6.1-sol reaches 0.568 against Opus's 0.546. Code, the task registry, our OS-level sandbox harness, and the live leaderboard are publicly available at https://github.com/Aayush-Marasini/WeightBench.
No comment yet.
Scooped by Philippe Gassmann
Today, 3:01 AM
Scoop.it!

[2611.18562] Training Reasoning Models in Simulation to Natively Use Real Robots

[2611.18562] Training Reasoning Models in Simulation to Natively Use Real Robots | AI News | Scoop.it
Robot-use agents (RUAs) have emerged as a promising approach to general-purpose robot control. However, strong performance has so far been achieved only by frontier proprietary models such as GPT-6 Astra and Claude Fable 5.2, whereas open-weight models remain substantially less capable. We show that an open-weight 27B reasoning model can be trained to use real robots natively through reinforcement learning, entirely in simulation and without demonstrations. Like a computer-use agent (CUA), the model observes the environment, reasons in a chain of thought, and generates the next end-effector pose, without any harness. We build a Post-training Engine for Agentic Robotics (PEAR) and use it to train the model on simulated robot tasks at scale with robot-use reinforcement learning (RURL), mirroring how frontier models acquire computer use. We find that the skills acquired in simulation generalize well beyond the training tasks: they transfer to an unseen robot, to the real world, and to novel instructions. After training on LIBERO and RoboDojo in simulation, the model succeeds in 60% of our trials on a real dual-arm robot, whereas the base model succeeds in none. On LIBERO, training raises the success rate from 1.6% to 74.4%, approaching GPT-6 Astra and GPT-6.1 Sol at maximum reasoning effort. Compared with $π_{0.5}$, a vision-language-action model fine-tuned on demonstrations, our model is substantially more robust to changes in the task itself. On LIBERO-Pro, $π_{0.5}$ succeeds in only 0.75% of episodes when assigned a new task in a familiar scene whereas our model succeeds in 57.4%. Finally, the model retains its general capabilities, with no measurable degradation on nine reasoning, coding and vision benchmarks. These results highlight robot-use reinforcement learning as a promising component of foundation-model post-training.
No comment yet.
Scooped by Philippe Gassmann
November 15, 2:02 AM
Scoop.it!

AI nella PA, dall'Europa al lavoro con il Politecnico di Milano

AI nella PA, dall'Europa al lavoro con il Politecnico di Milano | AI News | Scoop.it
Da qualche mese AI4Gov-X attiva esperti di 20 Paesi europei per formare funzionari pubblici in grado di guidare il cambiamento
No comment yet.
Scooped by Philippe Gassmann
November 15, 2:02 AM
Scoop.it!

Data center dell’IA più “verdi”: in arrivo bando Horizon Europe

Data center dell’IA più “verdi”: in arrivo bando Horizon Europe | AI News | Scoop.it
La Commissione europea, attraverso Horizon Europe, ha aperto il bando Energy efficiency and sustainability of AI data processing in Data Centres (IA) da 39
No comment yet.
Scooped by Philippe Gassmann
November 15, 2:02 AM
Scoop.it!

paul-nameless/agweb: A browser built for agents - no focus stealing, a tiny footprint, all tabs in one view. ·

paul-nameless/agweb: A browser built for agents - no focus stealing, a tiny footprint, all tabs in one view. · | AI News | Scoop.it
A browser built for agents - no focus stealing, a tiny footprint, all tabs in one view. - paul-nameless/agweb
No comment yet.