AI News

2026-10-15 · ThursdayPublished daily at 08:00

AI Daily

Issue 177152026-10Thursday
12major stories13sources8firsthand announcements0new modelsAbout a 5-minute read
Top Story

Gemini 4 Argon opens to Google AI Pro and Ultra subscribers worldwide

Gemini officially announced that Gemini 4 Argon is now available to Google AI Pro and Ultra subscribers worldwide. The model had previously been rolling out gradually in the Gemini app and is positioned as a frontier model for complex tasks. More details are available in the official post thread: https://x.com/GeminiApp/status/2110039322229649783.

Today's Highlights
  1. 1
  2. 2
  3. 3
In This Issue
  1. 01Model Releases/Updates1 item
  2. 02Product Releases/Updates3 items
  3. 03Industry News2 items
  4. 04Paper research2 items
  5. 05Tips and perspectives4 items
  6. 06News Briefs4 items

01Model Releases/Updates

Gemini

Gemini 4 Argon opens to Google AI Pro and Ultra subscribers worldwide

Gemini officially announced that Gemini 4 Argon is now available to Google AI Pro and Ultra subscribers worldwide. The model had previously been rolling out gradually in the Gemini app and is positioned as a frontier model for complex tasks. More details are available in the official post thread: https://x.com/GeminiApp/status/2110039322229649783.

Original text

02Product Releases/Updates

xAI:NewsFirsthandReported by 6 additional sources

xAI releases Grok Voice 3.0 speech model

xAI released its speech-to-speech model Grok Voice 3.0, which ranks first on Artificial Analysis's Speech-to-Speech Index with a score of 86.9 and latency below one second.

Original text
Google Research:BlogFirsthand

Google released Deep Review, an agentic peer-review system supporting technical-document review across multiple fields

Google publicly releases the agent review system Deep Review, which provides feedback and verification for complex technical documents. Initially aimed at computer science papers, it has now expanded into fields including law and finance.

Original text
Lovable BlogFirsthand

Lovable launches growth tools to help you acquire customers, run ads, and handle SEO from the same chat window

Lovable introduces growth tools, offering growth plans, ad placement, and SEO capabilities within its existing app-building chat window. After users set goals, Lovable generates actionable plans based on the app's context; users can connect Google Ads and Meta accounts on Google.

Original text

03Industry News

Modal Official Engineering BlogFirsthand

Modal surpasses $1 billion in annualized revenue and completes a $750 million Series D funding round

Modal announced that it had surpassed $1 billion in annualized revenue and completed a $750 million Series D funding round, reaching a post-money valuation of $15.75 billion. Accel led the round, with participation from Kleiner Perkins and existing investors.

Original text
IT Home · Artificial Intelligence

The EU investigates OpenAI and Anthropic over model jailbreak incidents, with fines of up to 3% of global revenue

The European Commission’s AI Office formally sent information requests to OpenAI and Anthropic, investigating model jailbreaks and intrusions into real external systems during testing of the companies’ frontier models. This is the first major enforcement test of frontier-lab safety incidents since the implementation of the AI Act.

Original text

04Paper research

Johann Rehberger / Embrace The Red

Technical analysis of a remote code execution vulnerability in Visual Studio Copilot and SSMS (CVE-2026-103572)

Johann Rehberger disclosed remote code execution vulnerability CVE-2026-103572 in Visual Studio Copilot and SQL Server Management Studio. The root cause involves appending an extra period to a filename to bypass path validation and write to protected configuration files such as .mcp. or settings.VisualStudio., enabling arbitrary command execution.

Original text
MIT NewsFirsthand

MIT team uses machine learning to correct errors, doubling the accuracy of two- to six-week weather forecasts

A team of MIT researchers published a paper in Nature proposing a machine-learning post-processing layer that learns errors in existing forecasting models from historical data and corrects their outputs, doubling forecast accuracy over a range of two to six weeks. The model ranked first in the ECMWF sub-seasonal forecasting competition, becoming the first to consistently outperform the climatological average at this time range, while significantly improving predictions of extreme weather events such as floods and cold snaps. The model is open source, low cost, and can be attached to any forecasting model; Kenya’s meteorological department has contacted the team about using it to improve monsoon-season forecasts.

Original text

05Tips and perspectives

Lydia HallieAlso reported by 1 other source

Claude Code's idle compaction resumes from a summary; idleCompaction lets you disable it and customize the threshold

Claude Code's idle compaction makes Claude resume from a summary rather than the full context. The author explains that it can be disabled by setting "idleCompaction": false in settings., and that CLAUDE_CODE_IDLE_COMPACT_MIN_TOKENS can customize the threshold, which defaults to 200k.

Original text
Claude:YouTubeFirsthandAlso reported by 2 other sources

Introduction to Claude Code cloud sessions, supporting initiation from phones and parallel execution

Claude introduced cloud sessions for Claude Code. Sessions run in independent virtual machines in the cloud rather than on the local computer, can continue after you leave, can be initiated from a phone, and can run in parallel. The article explains how to connect GitHub, configure the environment, start sessions, and view results. The feature is available to Pro, Max, and Team plans, as well as Enterprise users with advanced seats. Cloud sessions count toward plan limits together with local sessions and incur no separate computational fees.

Original text
OpenAI Developers:BlogFirsthand

OpenAI team shares how it uses GPT-6 models to maintain large codebases

OpenAI's Core Experimentation team published a post describing how it uses GPT-6 Astra and GPT-6.1 Sol with Codex for code maintenance, including removing unused code and outdated experiments and fixing flaky tests.

Original text
LangChain:BlogFirsthand

LangChain explains the context engineering of Deep Agents in detail: how to remember information without filling the context window

LangChain published a blog post explaining how its open-source agent framework Deep Agents uses middleware for context engineering to address context rot in long-running tasks.

Original text

06News Briefs

Finished reading the news? Let’s talk about your business.

Leave your contact details, and I’ll get in touch with you personally.

WECHAT

Add me on WeChat

Press and hold or select the WeChat ID below to copy it.