Mariusz

[Mariusz @MarMarRep]

Anthropic releases Claude Opus 4.8 - improved coding & agentic capabilities. Features dynamic workflows for Claude Code, effort control, & 3x cheaper fast mode. Same pricing as Opus 4.7. https://t.co/sEaC451HxI

[Mariusz @MarMarRep]

Smart Anthropic pattern: use mid-conversation system messages to switch AI orchestration modes without changing the main system prompt or breaking prompt caching.

Keeps latency + costs down while enabling deeper multi-agent workflows.

https://t.co/hEOXHxBPiB

[Mariusz @MarMarRep]

Anthropic released a free security-guidance plugin for Claude Code that autonomously reviews code edits for vulnerabilities in real time, reducing security-related PR comments by 30-40%. Available to all users.
https://t.co/ahwaagIIor

[Qwen @Alibaba_Qwen]

🙌🙌

[Mariusz @MarMarRep]

Google dropped Gemini 3.5 Flash, Managed Agents in the Gemini API, and Antigravity 2.0. The stack is getting more agent-first fast. https://t.co/zCYOrUWkQx

[Mariusz @MarMarRep]

Google I/O 2026: Gemini 3.5 Flash is now the default model for the Gemini app & Search AI Mode globally. New Gemini Spark is a 24/7 personal AI agent. Gemini Omni is a world model for video generation. All rolling out now. https://t.co/dTZGXjoDmY

[Mariusz @MarMarRep]

OpenAI released Codex in the ChatGPT mobile app, enabling remote development work from iOS/Android with live environment sync.
https://t.co/FJakiEwFob

[Claude @claudeai]

New in Claude Code: agent view.

One list of all your sessions, available today as a research preview.

[Mariusz @MarMarRep]

OpenAI launched the OpenAI Deployment Company: a majority-owned enterprise AI rollout arm with $4B+ initial investment and a planned Tomoro acquisition adding ~150 deployment specialists. Source: https://t.co/PvUKxzg7OZ

[Mariusz @MarMarRep]

Microsoft says global genAI use rose to 17.8% of the working-age population in Q1 2026, but the North/South adoption gap widened: 27.5% vs 15.4%. https://t.co/qiRRolLHQj

[Mariusz @MarMarRep]

Anthropic signed a SpaceX compute deal for Colossus 1: 300MW+ and 220k+ NVIDIA GPUs, enabling higher Claude Code and Claude Opus API limits. https://t.co/wfyugABVIQ

[Mariusz @MarMarRep]

Anthropic, Blackstone, H&F and Goldman Sachs are forming a new AI services company to bring Claude into midsized businesses, with Anthropic Applied AI engineers embedded on deployments. https://t.co/jPu0a8CPNK

[Mariusz @MarMarRep]

NVIDIA released Nemotron 3 Nano Omni, an open multimodal model for agents that handles video, audio, images and text in one system, claiming up to 9x higher throughput. https://t.co/mZkfnkVtir

[Mariusz @MarMarRep]

OpenAI open-sourced Symphony, a spec for orchestrating Codex-style agent workflows. It’s a notable push toward interoperable agent tooling rather than closed platform-only orchestration. https://t.co/JBEZckXLpO

[Mariusz @MarMarRep]

DeepSeek just open-sourced DeepSeek-V4 Preview: 1M context by default, new V4-Pro and V4-Flash, plus API support via OpenAI- and Anthropic-compatible endpoints. Source: https://t.co/LBNhqLiHzP

[Mariusz @MarMarRep]

OpenAI dropped GPT-5.5 today: rolling out in ChatGPT and Codex, with GPT-5.5 Pro for harder, higher-accuracy work. API access is “very soon.” https://t.co/YpS2SQf5hy

[Mariusz @MarMarRep]

Google launched Gemini Enterprise Agent Platform: a unified stack to build, govern and optimize AI agents, integrated with Gemini Enterprise and supporting Gemini plus Anthropic models. https://t.co/wTPyqZAYsM

[Mariusz @MarMarRep]

OpenAI launched Workspace Agents in ChatGPT: persistent agents for repeatable work across Slack, Salesforce, Notion and Atlassian Rovo. Research preview is live for Business, Edu, Enterprise and Teachers. https://t.co/SeKxMB8eiS

[Movez @0xMovez]

This 30-min workshop by the creator of Claude Code will teach you more about vibe-coding than 100 YouTube video guides.

Bookmark it & give it 30 minutes today. This video will change the way you use Claude forever.

[Mariusz @MarMarRep]

Anthropic launched Claude Opus 4.7 — its most capable model yet.

Better coding, stronger image analysis, improved instruction following, and more creative writing. Also adds new cybersecurity safeguards ahead of future Mythos models.

Source: https://t.co/6fLY3MfeT8