CORUSIA NEWS

कानून और सुरक्षा, नई तकनीक, रोबोटिक्स, ओपन सोर्स मॉडल और 1 मिनट से कम के AI वीडियो।

elon musk pledges to give one spacex share to every us child in 2027

Musk posted on X that he will give a SpaceX share to each US child in 2027. According to Genbeta, the estimate uses 72 million children and a $146.03 share price. The pledge lacks details on distribution and remains unverified by the company.

Genbeta

elon musk says tesla robotaxi outpaces waymo in austin for one week

Tesla logged 438,000 driverless miles in Austin in a single week, according to electrek. Waymo’s weekly average was 406,000 miles, based on data that is four months old. The article notes that Waymo still drives roughly seven times more driverless miles nationwide each week. Updated Waymo numbers remain unverified.

electrek.co

microsoft spends $190 b on ai and reorganizes business around agents

Microsoft has invested $190 billion in AI in 2026. The figure comes from Statista data cited by Xataka. The firm says AI agents are now its core organizational unit. This suggests Microsoft will steer revenue toward AI and infrastructure.

Xataka

joe tsai, alibaba president, backs open source ai for european sovereignty

Joe Tsai said open source AI is key for Europe. Alibaba reports that its Qwen3.8‑27B model is released under Apache 2.0. It matters because Europe relies on Chinese and US compute, says Xataka.

Xataka

Claude Duet brings Anthropic’s proactive assistance to small businesses

Europa Press reports that Anthropic has introduced Claude Duet, an assistant designed to take the initiative when helping small businesses. The report places it alongside Meta’s Muse and OpenAI’s Dots.

europapress.es

deepmind predicts hurricane isaías trajectory better than traditional models, hurri lab says

a deepmind model forecasted hurricane isaías track with an error of a few hours to three days, according to hurri lab analysis. it outperformed european and usgfs systems in locating the landfall in florida. however, the nhc noted its official forecast was 20‑40 % more skillful at predicting storm intensity. further validation is needed to see if the trajectory advantage holds for other storms.

Xataka

google cloud and nvidia announce general availability of vera rubin infrastructure

google cloud and nvidia make vera rubin instances generally available, according to the nvidia blog. availability begins next month and will expand to europe. this gives cloud customers a platform for large‑scale training and inference. exact rollout dates in Europe remain unconfirmed.

blogs.nvidia.com

elon musk says rival space plans lag behind starship

on october 17 2026 the ceo posted on x that competitors are falling behind. according to xataka, rocket lab delayed its neutron launch to 2027 due to a carbon‑fiber part defect. blue origin experienced a new glenn explosion in may, the company’s ceo reported. the setbacks may widen spacex's operational lead, but their long‑term effect remains unverified.

Xataka

Hans Anders pulls Meta Ray‑Ban smart glasses from over 400 stores

Hans Anders stopped selling Meta's Ray‑Ban smart glasses in its 400 stores across the Netherlands and Belgium, according to a company spokesperson. The withdrawal follows concerns about hidden recording and the indicator light, reported by the Offlimits association. The retailer will wait for regulatory clarification and possible measures before deciding on a relaunch. The exact number of units sold and the outcome of pending legal actions remain unknown.

Genbeta

unitree robotics unveils auto‑evolutionary system for humanoid robot training

unitree showcased in Hangzhou a project called "unitree rehearsal" that uses AI to retrieve research, generate control code and test robots, according to Xataka. founder Wang Xingxing said the aim is to speed up skill acquisition and cut human intervention, as Wang explained. the approach remains a proof of concept and its real‑world effectiveness is yet to be verified, according to the South China Morning Post.

Xataka

Jason Clinton, Anthropic's cyber security deputy, warns AI will worsen critical vulnerabilities

On October 22 Jason Clinton posted on X that artificial intelligence will make exploiting security flaws easier, according to his message. He referenced the recent PHP2Shell attacks rated 9.2 severity, reported by Reuters and confirmed by Amazon and Cloudflare. He notes that response teams face greater pressure and that patch deployment remains delayed. The warning draws on his background at Google and Anthropic, as described by Xataka.

Xataka

walmart launches ai agent beta while amazon blocks external access

walmart opened a beta allowing ai agents to shop for users, according to its CEO. amazon announced a ban on unauthorized agents, per its notice. the split highlights divergent retailer policies on customer data. a common standard for ai agents in e‑commerce is still missing.

Xataka

IBM says open-weight models deliver 90% of performance at 10% of cost

IBM states open-weight models reach 90 % of top-tier performance. IBM estimates token cost between 10 % and 25 % of closed models. If true, firms could cut AI expenses while keeping quality.

Xataka

Anthropic releases Claude Duet, proactive AI agent for professionals in the US

Anthropic has introduced Claude Duet, a proactive AI agent for professionals. According to Xataka, it monitors apps, sends morning summaries and stores memory in the cloud. Anthropic admits personal errands are not its strength. The launch matters as it shows competition in messaging‑integrated AI agents.

Xataka

FAHR-Face AI predicts mortality risk and biological age in cancer patients

A model called FAHR-Face was trained with 34,363 oncologic patient photos and validated on an independent cohort of about 5,000 patients, according to the study. The AI estimated biological age and classified mortality risk, finding that the high‑risk quartile had 3.23 times higher death probability than the low‑risk quartile. Among 122 high‑risk patients, 52.5 % died within 90 days, the data show. Researchers caution that these are statistical estimates and that prospective trials are still needed to prove clinical benefit.

Xataka

Google enables Gemini personal in Android Auto to access Gmail, Calendar and Maps

Google has activated Gemini Personal in Android Auto. According to the company, the assistant can query Gmail and Calendar. It also accesses Photos, Maps and Search while driving. The feature needs user consent, can be turned off, and is live in the US with broader rollout planned.

Xataka

Microsoft Research releases Flint-interactive, a language for interactive visualizations

Microsoft Research introduced Flint-interactive, a language that lets AI agents build interactive visualizations. According to the company's blog, users can click charts, filter data and highlight related content. This enables deeper data exploration beyond the author’s original view. Further testing is needed to confirm its practical impact.

microsoft.com

openai releases tool to check php vulnerability on websites

OpenAI released a tool to check a website's exposed PHP version. Xataka says phpcheck‑url sends a HEAD request to read server headers. It returns VULNERABLE, OK or UNKNOWN, but not a full audit. Useful for quick alerts, though UNKNOWN results need further investigation.

Xataka

agco to deploy agility digit humanoid robot in kansas factory

agco will install the agility digit robot at its kansas plant. according to the robot report, the pilot starts in 2027. the trial aims to test the robot's autonomy on an assembly line, but commercial‑scale performance remains unproven.

therobotreport.com

nvidia unveils RTX Spark and DGX Spark at gtc berlin

nvidia displayed its full local‑ai hardware lineup at gtc berlin. according to the blog, RTX Spark powers personal agents on windows PCs; DGX Spark lets developers build and scale applications on linux; DGX Station serves enterprise workloads on linux and will launch a windows version in december. the rollout expands local offerings, though pricing and availability remain unconfirmed.

blogs.nvidia.com

✒️ CORUSIA के योगदानकर्ता

⚡ Shorts, Reels और TikToks (< 1 मिनट)

AI से बने हमारे क्लिप और मूल प्लेटफ़ॉर्म से जुड़े क्रिएटर वीडियो।

YouTube · TikTok · Instagram

🎧 पॉडकास्ट और यूट्यूब वीडियो (HI)

हिंदी और हिंग्लिश में चुनिंदा टेक चर्चाएं।

YouTube · Spotify