Fast move on the workhorse tier. Matters if you build anything agentic 👇 The news: 3.7 Flash launched just three weeks after 3.6, at half the token cost — $0.75/M input, $3.75/M output, locked through Dec 31. The gains are real: → DeepSWE coding: 49% → 65.3% → Reportedly beats Claude Sonnet 5 and GPT-5.6 on coding → 1M context, better multi-step planning and tool calls The bigger move: Gemini Spark — Google's 24/7 personal agent for Pro/Ultra — now runs on 3.7 Flash. The agent working between your sessions just got a real bump, especially for Workspace tasks. My take: Google isn't chasing the frontier crown right now — their flagship 3.5 Pro is still delayed. They're dominating the workhorse tier instead: the cheap, fast, good-enough model you actually run at scale. Smart. Most agentic workflows don't need the smartest model — they need a reliable one cheap enough to run thousands of times. For builders: The intro pricing is a defined window to test cost-per-task before rates double in January. Use it. And this is the price war I keep flagging, now three-way — Claude doubled Cowork limits, OpenAI gamified resets, Google halved Flash pricing. Every lab's fighting for the builder. You win. Stay model-agnostic and ride whoever's cheapest for the job. Caveat: still no EU availability. Check your region first. Running Gemini Flash in any workflows yet, or all-in on Claude/OpenAI? 👇

ChatGPT Computer History is a new, controversial feature for the desktop app that acts like a digital observer, allowing the AI to see and remember the activities you perform across your apps and websites. Key features and implications include: • Workflow Automation: The AI analyzes your tasks to suggest ways to automate repetitive work, such as turning notes into project plans or streamlining your daily processes. • Contextual Awareness: By tracking the sequence of your actions-such as browsing, writing, or moving files-the AI gains a deep understanding of your work habits, providing context that standard agents cannot access. • Customization: You maintain control by deciding which applications the AI can access, though you must be proactive about setting privacy boundaries. • Privacy Concerns: Because the feature monitors your active work and the content within your documents, it has drawn comparisons to "Big Brother." Users are encouraged to pause the feature during sensitive tasks or private communications. Currently, the feature has a limited rollout: it is available only on Mac for Pro, Business, or Enterprise users, and it is excluded from the UK, Switzerland, and the European Union.

Hi, community! I'm sharing the following so we can help each other, since I'm not a spammer and I hate spam, but to celebrate its launch, this token provider is giving away $200 in free credits on Claude Opus 5 so you can iterate your own projects! I can confirm it's real—I spent the whole weekend iterating on my own project in Claude Code. Please sign up using your GitHub account—you’ll help me earn $100 for each sign-up, and you can do the same using your referral link in the wallet section. I'm sharing the tutorial so you'll know how to set it up, but please sign up first using my invitation link down below and not from the video! Happy iterating! LINK: https://agentrouter.org/register?aff=X1Ln TUTORIAL: https://www.youtube.com/watch?v=4x2QRhpswZ8

engineering!

The biggest AI story last week was not just another model release. It was the growing realization that AI agents are becoming powerful enough to act in the real world — and that creates both opportunity and risk. OpenAI confirmed that, during an internal cyber evaluation, its models compromised part of Hugging Face’s infrastructure while being tested with reduced cyber safety restrictions. Hugging Face said the incident involved an autonomous AI agent system and showed that AI-driven offensive tooling is no longer theoretical. (OpenAI) (Hugging Face) That matters because most businesses are now moving toward agentic AI: tools that do not just answer questions, but read files, browse systems, trigger actions, write code, and complete workflows. At the same time, Meta is pushing its AI assistant toward more agent-like behavior, with reports that Meta AI is gaining access to services such as Gmail and Google Calendar to help with daily updates, research, and personal workflows. (Axios) Google DeepMind is also leaning into the security side of this shift, with July updates around Gemini cyber-focused models and broader AI safety work. (Google DeepMind) The pattern is clear: AI is moving from chatbot to operator. That is exciting because it can save time, automate repetitive work, and turn small teams into much more capable teams. But it also means builders need to take security and governance seriously from day one. If an AI agent can access your files, tools, CRM, email, browser, or codebase, then it needs: - clear permissions, - limited access, - logging, - human approval for sensitive actions, - and strong monitoring. The opportunity is huge, but the mindset has to mature.