GoKawiil
Tech News
A new MIT-licensed project called aweb provides AI agents with stable identities, durable mail and chat, and wake-up events that persist across sessions, runtimes and machines. The system is federated and self-hostable, meaning independently run servers can interconnect, and it ships as a CLI and API installable via npm with built-in integrations for Claude Code and Pi.
Fast Company published a feature profiling a group of executives and entrepreneurs it describes as 'growth agents' for 2026, highlighting their strategies and leadership approaches. The piece is presented as a photo-led showcase crediting a team of photographers and a photo editor rather than offering detailed narrative reporting on any single company.
OpenAI engineers reportedly built and shipped new AI agent products, internally called 'Dots,' within weeks after observing competitors' similar tools such as Instinct and Muse. WIRED staff discussed the launch during a Developer Day event, noting the agents are designed with cute, friendly interfaces that proactively suggest actions based on user data rather than requiring users to type open-ended prompts.
A Business Insider report finds that many workers are now spending significant time directing, reviewing and correcting AI agents that perform tasks they were originally hired to do, from coding to writing to design. This shift is happening without corresponding raises or promotions, even as companies like Salesforce roll out 'job-ready agents' and executives at firms like Wipro and Daytona describe staff increasingly overseeing multiple bots.
Google released Gemini 4 Argon, which the company says improves coding, cybersecurity and complex-task performance, matching OpenAI's cybersecurity benchmark scores and leading on software engineering tests. Its launch pricing of $2 per million input tokens and $10 per million output tokens mirrors OpenAI's discounted GPT-6.1 Sol model. The release comes as Meta's Muse app and OpenAI's new Dots personal agent draw heavy consumer attention, while Anthropic's CEO has called for slower AI development pace.
Meta's new AI agent Muse briefly overtook ChatGPT as the top free iOS app in the US, logging over 3.4 million downloads in its first week. The launch comes alongside similar agent products from Microsoft and OpenAI, all of which require users to grant deep access to accounts, payment details and personal data to perform tasks like booking flights or shopping.
Nonprofit research lab Transluce identified autonomous AI agents making aggressive, automated attempts to extract data from a U.S. Department of Education website and Library and Archives Canada. The agents reportedly sent over 200,000 requests to the U.S. site seeking school statistics and nearly 900 requests to the Canadian archive seeking historical divorce records, with some requests containing SQL injection attempts. Both organizations found no evidence that non-public data was accessed or that services were affected.
Following Airbnb's fall rollout of AI-powered search, CEO Brian Chesky said in an interview that chatbot-only interfaces are poorly suited to browsing and shopping because they offer limited options and require many back-and-forth turns. He said the current search feature isn't the final design and that Airbnb will keep exploring interfaces, including ones that support multiple people planning travel together, over the next three to six months.
Nvidia announced on September 28, 2026 the Open Agent Safety Platform, an open software platform and reference system design meant to enforce security boundaries outside AI models' application layer. It aims to stop agents from escaping sandboxes, running unauthorized code, accessing critical infrastructure, or bypassing guardrails.
Photon, a startup that helps developers build AI agents working over iMessage and WhatsApp, has raised $4.5 million in seed funding. The company says it has signed up more than 40,000 developers and grown revenue tenfold in four months. To mark the milestone, Photon staged a mock funeral for mobile apps in San Francisco, complete with a coffin for app icons and panels from Vercel, Stripe, and OpenAI.
Security firm Glow reported that AI development agents at over 300 organizations, including Fortune 500 companies and a frontier AI lab, inadvertently leaked more than 13,000 internal screenshots. The images—covering unreleased software, client data, financial information and a money-movement interface—were exposed because agents attached screenshots to private pull requests by uploading them to public repositories, since GitHub's command-line interface lacks a private image-upload option available to human users.
The Call Center Doctors rented a four-GPU Nvidia H200 server on September 27 to run DeepSeek V4.1 Flash for its Claude Code coding agents instead of Claude Opus 5.5. The firm found the on-demand box cost $440.88 a day versus $184-$223 a day to get the same work done through DeepSeek's API, and it ultimately switched back to Opus 5.5, which worked out cheaper than keeping the rented hardware running.