Keldura Daily Open Keldura

Keldura Daily · AI & Technology

AI-found PHP flaws prompt website shutdowns and emergency patching

Governments and companies shut down websites while responding to two critical flaws in PHP, a programming language used by about 70 percent of websites, according to Wiz Research. Etsy announced rules requiring listings to include an unedited photo of the actual item against a real background, with the first photo unedited. OpenAI released the Decisions API for bounded-answer tasks such as classification, routing and choosing a next action, claiming speeds up to 10 times faster than GPT‑6 Luna through the Responses API. Uber announced a rollout of ChatGPT Work to more than 30,000 employees, expanding an initial 1,000-person pilot into its first company-wide agentic platform for knowledge work.

The field note

2 sources · 2 items
  1. Amazon Threat Intelligence identified a hacking attempt before public disclosure, but said it was unclear how t…
  2. Calif researcher Ryan Austin said exploitation did not appear widespread at the time of reporting; discovering…
  3. Many large platforms stayed online because they used other languages or had operators able to install the fix q…
Story 011 source

AI-found PHP flaws prompt website shutdowns and emergency patching

Governments and companies shut down websites while responding to two critical flaws in PHP, a programming language used by about 70 percent of websites, according to Wiz Research. PHP programmers released a fix on October 21, and authorities urged organizations to patch immediately or take vulnerable systems offline. [5] Security firm Calif discovered the flaws in August using OpenAI's GPT-5.6 Sol during an OpenAI-funded audit. [5]

Why it matters

AI-assisted discovery can expose weaknesses in widely used software faster than operators can necessarily repair them. The exposure is narrower than PHP's overall footprint: AFP reports that only sites running PHP on their own servers are affected. [5]

Key insights

  • Amazon Threat Intelligence identified a hacking attempt before public disclosure, but said it was unclear how the actor obtained exploit information. [5]
  • Calif researcher Ryan Austin said exploitation did not appear widespread at the time of reporting; discovering a severe flaw is not the same as establishing widespread compromise. [5]
  • Many large platforms stayed online because they used other languages or had operators able to install the fix quickly. [5]
Story 021 source

Uber expands ChatGPT Work from pilot to company-wide deployment

Uber announced a rollout of ChatGPT Work to more than 30,000 employees, expanding an initial 1,000-person pilot into its first company-wide agentic platform for knowledge work. [9] The platform connects to tools including Google Drive, Gmail, Google Calendar and Slack to support tasks such as research, meeting preparation, presentations and cross-application workflows. [9]

Why it matters

This moves agents from a limited pilot into broad workplace availability, where they can work across connected business applications rather than only answer questions. The announcement establishes deployment scale and intended capabilities, but provides no measured productivity or quality results. [9]

Key insights

  • The announced employee reach is more than 30 times the size of the initial pilot. [9]
  • OpenAI says GPT‑6.1 Sol will be the default capability within ChatGPT Work, alongside voice and image-generation options. [9]
  • The platform is designed to complete multi-step tasks across applications, making its intended role broader than a standalone conversational assistant. [9]

Create your own daily briefing — start free. Keldura monitors the sources you choose and gives you a private, grounded daily digest with cited answers.

Create your own daily briefing — start free