AI News
2.4K views | +0 today
 
Scooped by Philippe Gassmann
onto AI News
Today, 2:02 AM
Scoop.it!

keepplain/keepplain-plugin ·

keepplain/keepplain-plugin · | AI News | Scoop.it
Contribute to keepplain/keepplain-plugin development by creating an account on GitHub.
No comment yet.
Your new post is loading...
Your new post is loading...
Scooped by Philippe Gassmann
Today, 3:01 AM
Scoop.it!

SoftBank chief approached lenders for SK Hynix acquisition before dropping plan | Asianomist

SoftBank chief approached lenders for SK Hynix acquisition before dropping plan | Asianomist | AI News | Scoop.it
Masayoshi Son contacted Wall Street institutions about financing a potential deal valued at roughly US$1 trillion, though the company says the proposal is no longer active.
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

Nasce la fiera dell’intelligenza artificiale governata dall’uomo - Le Cronache Attualità

Nasce la fiera dell’intelligenza artificiale governata dall’uomo - Le Cronache Attualità | AI News | Scoop.it
Comincia dall’Università degli Studi di Salerno la strada verso la terza edizione dello Human Driven AI SummitTM, in programma a Pompei il 16 e 17 aprile 2027. Il primo capitolo di Road to Pompei, “Human Driven AI Policy”, si è tenuto il 19 novembre nell’Aula G. De Rosa del campus di Fisciano e ha portato […]
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

keepplain/keepplain-plugin ·

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

Bot Auto launches Capacity Grid for 200 million miles of driverless trucking

Bot Auto launches Capacity Grid for 200 million miles of driverless trucking | AI News | Scoop.it
Bot Auto has launched its Capacity Grid, a freight transportation network designed to match customer demand with autonomous trucking capacity, alongside agreements covering 200 million miles of dri…
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

Phison Previews an SSD That Runs AI Frameworks on Its Own NPU

Phison Previews an SSD That Runs AI Frameworks on Its Own NPU | AI News | Scoop.it
Everyone's going all in on AI, so SSD manufacturers were not going to be left behind. Phison has detailed an enterprise SSD design that can run AI frameworks on the drive itself. At the Supercomputing 2026 (SC26) event in Chicago, the company showed off the computational storage design tha
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

[2611.23219] Do LLM Agents Still Need to Be Told How to Self-Improve?

[2611.23219] Do LLM Agents Still Need to Be Told How to Self-Improve? | AI News | Scoop.it
LLM agents increasingly drive their own learning. Self-improvement methods delegate various components of model optimization to the models themselves. As models become more capable, this motivates a basic question: do these increasingly capable models still need us to specify how they should self-improve? In this paper we study an autonomous paradigm in which, given basic primitives, resources, and a general goal for self-improvement, agents construct their own self-learning algorithm for context-based optimization. Using the setting of gameplay as a testbed, a frozen LLM agent is given a compute budget and tools for collecting experience and updating text documents, and allowed to freely experiment without an externally imposed loop or access to real-world data. We find that sufficiently capable models do self-improve under these conditions, demonstrating strategies that align with classic cognitive frameworks for self-regulated learning and constructs from machine learning literature on self-generated curricula. Furthermore, we study the extent to which the learned text artifacts transfer to external benchmarks, observing improvement across different game horizons, mechanics, and 3D navigation.
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

[2611.23143] TCP4D: A Unified Tool-Center-Point Tracking and Generation Model for Robot Manipulation

[2611.23143] TCP4D: A Unified Tool-Center-Point Tracking and Generation Model for Robot Manipulation | AI News | Scoop.it
Robot manipulation videos provide a rich visual experience, yet recovering robot action and gripper-state annotations from RGB observations still remains challenging. Using the embodiment-agnostic tool-center-point (TCP) pose as the action space, we present TCP4D, a unified model in which a shared 4D scene representation both tracks observed TCP motion and generates future TCP trajectories. Given a monocular video and initial 2D TCP queries, the tracker directly recovers metric TCP poses in SE(3) and gripper states alongside scene geometry and camera parameters. The generator samples local features from the same 4D representation along the recovered TCP tracks, then combines them with robot observations and language instructions to predict future TCP states. Our new cross-embodiment augmentation pipeline converts existing RGB videos into demonstrations for alternative embodiments by retargeting the recovered motion and synthesizing the corresponding observations, without recorded source-robot states. Experiments show that the tracker achieves absolute trajectory errors of 7.60 and 16.07 mm on DROID and RoboLab and adapts to a new embodiment with only ten demonstrations. The generator achieves 69.19% mean success on RoboTwin and outperforms well-established policy models on two real-world tasks, while joint training on source and augmented demonstrations improves target-robot success by 41.5% over augmented-only training.
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

[2611.21099] ForceHarness: Turning Force into Action and Physics for Frozen Frontier Models

[2611.21099] ForceHarness: Turning Force into Action and Physics for Frozen Frontier Models | AI News | Scoop.it
Frontier multimodal models can now control robots zero-shot from images, yet many manipulation tasks depend on contact forces that images cannot show. Bringing force to an API-only frozen model raises two challenges. First, contact evolves much faster than an API model can decide, so the system must close the loop on force beneath the model. Second, a frozen model cannot interpret raw force streams, so force must be returned in a form the model can reason about. Therefore, we present ForceHarness, a harness that lets frozen models use force without demonstrations or finetuning. After calibrating hand force from a wrist sensor or joint torques alone, ForceHarness addresses the first challenge with self-terminating force primitives that regulate force on their own and return at a contact event. To address the second challenge, a physics translation layer turns each force trace into physical quantities with uncertainty as the tool result. To evaluate ForceHarness against expert force controllers and learned policies under identical conditions, we also build HiddenPhys, a seven-task MuJoCo benchmark with hidden physical variables. On a real arm without a wrist force sensor, ForceHarness raises the zero-shot success rate of GPT-6 Astra from 3.3% to 90.0%. In simulation, it adds 17.5 percentage points on the public ForceBench suite and 34.3 percentage points on HiddenPhys, surpassing force-aware policies trained on demonstrations on both benchmarks and matching hand-written expert controllers on HiddenPhys. Ablations further show that smaller models make significantly better decisions when given physical quantities rather than raw force values.
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

Not for gamers: Nvidia's new 50,000-core Rubin HPC chip dwarfs the RTX 5090 — yet independent benchmarks won't prove its real-world speed until 2028

Not for gamers: Nvidia's new 50,000-core Rubin HPC chip dwarfs the RTX 5090 — yet independent benchmarks won't prove its real-world speed until 2028 | AI News | Scoop.it
Nvidia reveals a monster 50,000-core GPU to power AI science agents — joining its Vera Rubin supercomputing platform with seventh-gen NVLink in 2028
No comment yet.
Scooped by Philippe Gassmann
Today, 2:01 AM
Scoop.it!

Amazon nears $7.5B deal to buy AI start-up Decart (AMZN:NASDAQ)

Amazon nears $7.5B deal to buy AI start-up Decart (AMZN:NASDAQ) | AI News | Scoop.it
Amazon (AMZN) is in final talks to acquire AI startup Decart for about $7.5B, with a deal possible as early as next week but still at risk of collapse. Read more here.
No comment yet.
Scooped by Philippe Gassmann
November 21, 3:02 AM
Scoop.it!

Archer raises $10M to connect recruiters and job seekers with AI

Archer raises $10M to connect recruiters and job seekers with AI | AI News | Scoop.it
Archer develops an AI-powered recruitment platform that automates candidate sourcing and qualification, helping employers identify suitable talent and job seekers discover relevant opportunities.
No comment yet.
Scooped by Philippe Gassmann
November 21, 3:02 AM
Scoop.it!

Ex-Alibaba Star’s AI Startup Doubles Valuation to $5 Billion - Bloomberg

Ex-Alibaba Star’s AI Startup Doubles Valuation to $5 Billion - Bloomberg | AI News | Scoop.it
A startup founded by Alibaba Group Holding Ltd.’s former AI chief has closed new funding at a valuation of $5 billion, more than doubling its price tag in mere months in a positive sign for smaller developers trying to squeeze into a crowded artificial intelligence arena.
No comment yet.
Scooped by Philippe Gassmann
November 21, 3:02 AM
Scoop.it!

CDP Venture Capital, in arrivo un direttore generale: dopo lo stallo di mesi, la sfida della nuova governance

CDP Venture Capital, in arrivo un direttore generale: dopo lo stallo di mesi, la sfida della nuova governance | AI News | Scoop.it
L'ipotesi più accreditata è quella di Stefano Piccaluga, manager Invitalia, come codirettore al fianco dell'AD Emanuele Levi. Ma non c'è ancora la quadra politica. Resta aperto il tema della governance in una società che ha cambiato tre volte il vertice in sei anni. Mentre si avvicina il momento di decidere il futuro della rete degli acceleratori
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

Le donne che guidano le startup deeptech italiane | RDD

Le donne che guidano le startup deeptech italiane | RDD | AI News | Scoop.it
Le donne guidano il 13,84% delle startup innovative italiane. Scopri chi sono le founder nel deeptech, come accedono ai finanziamenti e quali barriere affrontano.
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

Nuovi materiali per sostituire le terre rare: Rara Factory vince Startup Marathon 2026

Nuovi materiali per sostituire le terre rare: Rara Factory vince Startup Marathon 2026 | AI News | Scoop.it
Primo posto alla startup venezana nella finale della settima edizione del contest promosso da Area Science Park, UniCredit Start Lab e Fondazione Comunica ospitata da UniCredit a Milano. Al secondo posto la triestina Mirron
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

Show HN: Zero C# – C# without a runtime or GC | Hacker News

No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

Inference drives two-thirds of AI workloads, changing chip deployment

Inference drives two-thirds of AI workloads, changing chip deployment | AI News | Scoop.it
Modular data centers fit Positron AI Inc.'s low-power inference servers, which use commodity memory and let operators deploy in 90 to 120 days.Inference is now two-thirds of AI workloads, and modular data centers can go from land and power to tokens in 90 to 120 days, says Positron AI.
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

Medtronic's Hugo surgical robot ready for bigger operation

Medtronic's Hugo surgical robot ready for bigger operation | AI News | Scoop.it
Medtronic's Hugo surgical robot received FDA clearance for general surgery, expanding its U.S. reach as the company pursues new applications.
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

[2611.23034] M3OPD-Reasoner: An Open Recipe for Post-Training Multimodal Reasoning with Multi-Teacher On-Policy Distillation

[2611.23034] M3OPD-Reasoner: An Open Recipe for Post-Training Multimodal Reasoning with Multi-Teacher On-Policy Distillation | AI News | Scoop.it
Multi-Teacher On-Policy Distillation (MOPD) enables a single student to absorb the skills of several domain teachers by learning from its own generated responses. We present Multimodal Model Multi-Teacher On-Policy Distillation Reasoner (M3OPD-Reasoner), a fully open recipe for post-training vision-language models (VLMs) over three reasoning domains: image STEM, image perception, and video. We curate two datasets, M3OPD-SFT-1.35M for supervised fine-tuning (SFT) and M3OPD-RL-193K for reinforcement learning (RL) with verifiable rewards. On these data, we compare post-training recipes that start from Qwen3.5-9B: training one model on the mixed data of all domains (MixSFT and MixRL), and training one teacher per domain with Group Sequence Policy Optimization (GSPO), directly or after an SFT warm-up, and distilling the teachers into one student with MOPD. We evaluate on 21 image STEM, image perception, and video reasoning benchmarks after we decontaminate the training data against them. The comparison establishes MOPD of teachers trained with GSPO directly from Qwen3.5-9B as the best recipe: the student closely matches each teacher on its own domain and is on average above the mixed-training baselines on the same prompts. Since MOPD already merges the domain capabilities that MixRL would otherwise have to learn, its student is a stronger starting point for RL, and we ask whether the same data can carry the model further from there: we add MixRL on top of the MOPD student, on the same prompts and with the same settings, which, to our knowledge, no prior VLM recipe has studied. The same prompts still improve the MOPD student, with no collapse, and give our main model, M3OPD-Reasoner-9B. MOPD and MixRL can therefore work as consecutive stages that get more out of the same data. We release all data, the training code, and the checkpoints of every stage.
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

[2611.21392] VideoNexus: Video Perception Beyond the Reasoning Context

[2611.21392] VideoNexus: Video Perception Beyond the Reasoning Context | AI News | Scoop.it
Understanding long videos requires locating evidence, resolving visual details, and integrating observations into an answer. Even when relevant evidence is present, performing these operations within long, dense visual sequences burdens both perception and reasoning. Video agents enable selective inspection, but returning frames to the planner still accumulates visual information in its reasoning context. Delegating perception reduces this burden, yet sparse sampling can miss decisive details, and per-frame descriptions can preserve context growth. We introduce VideoNexus, a training-free framework that separates detailed video perception from evidence-driven planning and reasoning. The planner controls where to inspect, what to resolve, and how densely to sample. Perception sub-agents process native audio-video in independent contexts and return compact, temporally grounded evidence for iterative search and verification. Using Gemini 3.8 Flash for both planning and perception, VideoNexus achieves 88.6% accuracy on full LVBench, exceeding same-backbone static ingestion and built-in agentic processing by 3.6 and 1.1 points, respectively, at 44% of the latter's measured API cost. A paired LVBench subset ablation shows lower planner input and higher accuracy. With Kimi K3 as the shared backbone, VideoNexus improves over the official Video-GAIA ReAct harness by 8.5 points in our evaluation. These results support compact evidence exchange for combining detailed observation with efficient planning and reasoning.
No comment yet.
Scooped by Philippe Gassmann
Today, 2:02 AM
Scoop.it!

Rapidus reportedly struggles to produce working SRAM on its 2nm node, putting 2027 mass production at risk — company faces significant challenges using EUV to develop GAA transistors

Rapidus reportedly struggles to produce working SRAM on its 2nm node, putting 2027 mass production at risk — company faces significant challenges using EUV to develop GAA transistors | AI News | Scoop.it
Japan's semiconductor startup has demonstrated functioning GAA transistors, but integrating them into working memory cells appears to be a considerably greater challenge.
No comment yet.
Scooped by Philippe Gassmann
Today, 2:01 AM
Scoop.it!

宝玉 on X: "2026 年 11 月 20 日,Anthropic 的数学家 Levent Alpöge 宣布,他和 Claude 一起证明了黎曼猜想,交出来的是 1700 万行 Lean 代码。
这期用 5 分钟讲清楚:
黎曼猜想说的是什么,和质数有什么关系;
为什么 160 多年没人证出来;
AI 在里面做了什么;
机器检查全部通过,算不算真的证明了;
银行加密会不会因此失效。
...

2026 年 11 月 20 日,Anthropic 的数学家 Levent Alpöge 宣布,他和 Claude 一起证明了黎曼猜想,交出来的是 1700 万行 Lean 代码。

这期用 5 分钟讲清楚:
黎曼猜想说的是什么,和质数有什么关系;
为什么 160 多年没人证出来;
AI 在里面做了什么;
机器检查全部通过,算不算真的证明了;
银行加密会不会因此失效。

同一天,斯坦福也放出一份纯人工证明草稿,几小时后就有人借助 AI 报告了问题。

两边对照下来:机器说没错,人还没读懂;人写的,正在大改。
No comment yet.
Scooped by Philippe Gassmann
November 21, 3:02 AM
Scoop.it!

Irish-founded software start-up Hoplite raises $5.5m in funding –

Irish-founded software start-up Hoplite raises $5.5m in funding – | AI News | Scoop.it
Platform provides easier way for developers to run AI coding agents in the cloud
No comment yet.
Scooped by Philippe Gassmann
November 21, 3:02 AM
Scoop.it!

Indian EO Startup SatSure Raises $21M, With Sovereign Fund Investment

Indian EO Startup SatSure Raises $21M, With Sovereign Fund Investment | AI News | Scoop.it
Earth Observation company SatSure raised a $21 million Series B round in a mix of equity, debt, and secondary transactions, with investment from India’s
No comment yet.
Scooped by Philippe Gassmann
November 21, 3:02 AM
Scoop.it!

AI Funding News November 20 2026: DeepSeek Leads $15.07B Across Six Rounds

AI Funding News November 20 2026: DeepSeek Leads $15.07B Across Six Rounds | AI News | Scoop.it
AI funding news November 20 2026: six AI startups raised about $15.07 billion in disclosed amounts, and DeepSeek accounted for more than 99% of it.
No comment yet.