
Welcome to todayās edition of AI News Orb
Sam Altman questions local RAM use, Claude controls a mouse brain experiment, Chainalysis confirms a cyberattack, and Claude helps users find forgotten subscriptions and reduce wasted spending.
Latest AI News
Trending AI, SaaS & Cybersecurity tools
AI & Technology News
Cybersecurity News
AI Lesson
AI Art Generation
Todayās Resource
Recommended AI Video
Latest AI News
Open AI
Sam Altman Questions RAM Use in Home PCs

Source: tomās Hardware
OpenAI CEO Sam Altman has criticized the idea of people running AI models on their own computers. In an interview with the YouTube channel Cheeky Pint, he called local AI processing āquite terribleā from a global resource-allocation perspective.
The Details:
Altman argued that large amounts of computer memory (DRAM) sit unused in personal computers for much of the time. He believes cloud data centers can use these resources more efficiently.
Altman's comments highlight a growing debate about where AI should run: on personal computers or in cloud data centers. His position favors centralized computing, where AI companies can manage and share computing resources across their systems.
However, it also raises questions about the future of running AI models locally.
The debate between local and cloud-based AI is worth following. The key question is whether personal computers will remain important for running AI models as data centers expand.
Anthropic
Claude AI Performs Brain Surgery on a Mouse

An AI agent powered by Anthropic's Claude Sonnet 5 independently selected a gene therapy, brain region, and injection coordinates for a mouse with experimentally induced epilepsy. It then controlled a motorized surgical device to deliver the treatment.
The Details:
The experiment was led by Gabriel Ocana-Santero, co-founder and CEO of Exin Therapeutics. The AI received a list of 19 possible gene therapies and one instruction: reduce the mouse's seizures.
This experiment shows how AI agents could take a more active role in scientific research, moving beyond analyzing data to making treatment decisions and controlling laboratory equipment.
However, it also highlights the challenges of trusting AI with complex, safety-critical tasks.
The mouse reportedly recovered well, but the findings have not been peer-reviewed. A technical report is expected. Further research will show whether this approach can work reliably in other experiments.
Trending AI, SaaS & Cybersecurity tools
Jotform lets users create apps directly through ChatGPT and Claude conversations.
Mistral AI released Mistral Large 4 with open weights. It supports reasoning, coding, and image analysis.
Canva unveiled Brush, an AI model that edits existing designs using simple instructions. It makes design editing faster and easier.
Aikido offers an AI-powered cybersecurity platform for code, cloud, and applications. It helps developers detect and fix security vulnerabilities faster.
Chainalysis
Chainalysis Suffers AI-Assisted Cyberattack

Blockchain analytics company Chainalysis has confirmed a cyberattack in which an external attacker accessed its internal systems and tools. The company said the attacker used what it described as āoffensive agentic techniques.ā
The Details:
The incident exposed some customer information, including user IDs, full names, email addresses, and organizational roles.
Chainalysis said it found no evidence that investigation data, authentication credentials, or session tokens were accessed.
The incident highlights a growing security challenge: attackers may use AI-powered tools to break into company systems.
AI can help make cyberattacks faster and easier to scale, creating new challenges for security teams.
Chainalysis is continuing its security response. The company has not disclosed the exact attack methods or how many customers were affected. Further updates should clarify the full impact of the incident.
AI & Technology News
Google DeepMindās PlanAssist cuts planning assessment time by 58% in early UK trials. AI could reduce paperwork and speed up home improvement approvals.
US Senators Ron Wyden and Edward Markey are investigating privacy risks linked to Meta's AI glasses. Reports raise concerns about secret recordings and sensitive footage being reviewed by contractors.
Privacy-focused AI assistant Ollie will shut down on November 6, 2026. The closure highlights the challenges of competing in the crowded AI assistant market.
Cybersecurity News
Hackers stole $6.36 billion. North Korean hackers are suspected.
10 blockchains affected; $3 billion frozen. Binance continues its investigation.
Cloudflare mitigated 18.2 million network-layer DDoS attacks in Q3 2026. The growing number of attacks highlights rising threats to online services.
Eye Security patched an abandoned Popcorn Time updater that put millions of devices at risk. Attackers could exploit abandoned domains to deliver malicious software.
AI Lesson
Find Wasted Subscriptions With Claude AI

A new AI workflow shows how to use Claude and Gmail to find forgotten subscriptions and track recurring payments automatically.
Many people pay for software and online services they rarely use. This workflow helps identify unnecessary spending, duplicate tools, and unexpected price increases.
Create a Gmail label called "Receipts."
Filter subscription emails from Stripe, PayPal, Apple, Google Play, and other services.
Connect Gmail to Claude.
Ask Claude to build a subscription tracker with vendor names, prices, billing dates, and payment cycles.
Schedule a weekly audit to identify new charges, price increases, and duplicate subscriptions.
Review Claude's report, including your monthly subscription total.
Claude can organize and flag suspicious spending, but you decide which subscriptions to cancel. Regular reviews can help prevent forgotten payments from adding up.
AI Art Generation
Create a similar photo using the Prompt Below

Prompt:
A side-by-side split screen image. Left side: A blue [Pepsi] aluminum soda can sliced into floating segments with dark soda splashing out and flying [ice cubes], set against a solid [blue] background. Right side: A green [Sprite] soda can sliced into floating segments with a clear splash and floating [lemon and lime slices], set against a solid [lime green] background. Commercial advertising photography style with dynamic splash effects and high-speed action capture.
Todayās Resource
If you have a Gmail account, read this

Google quietly connected your Gmail to Gemini AI.
Most people still donāt know the default settings let Gemini process your emails, attachments & data.
A class-action lawsuit (Thele v. Google) is ongoing over this. Google says it doesnāt train models on your Gmail, but the AI still gets access for summaries, replies & personalization.
Hereās the exact way to limit it
