Author: IBL News

IBL News | New York

OpenAI rolled out this week ChatGPT Images 2.0, a new tool that allows users to search the web and create visual explainers based on uploaded files, reasoning through the structure of the image before generating it.

According to the company, this image generator creates more sophisticated images, with improvements in its ability to follow instructions and preserve the details the user chooses.

It creates up to 8 images at once with thinking enabled, all while maintaining the same characters, objects, and styles across scenes.

OpenAI said this should make it easier to generate things like manga pages, a series of social graphics, or design plans for every room in a house.

It generates images with a resolution of up to 2K and in a range of aspect ratios, from wider formats such as 3:1 to taller ones like 1:3.

OpenAI first released ChatGPT Images last year and launched its latest major update in December, adding faster image generation and improved photo-editing capabilities.

Since then, competition has only been getting stronger, with the arrival of tools like Google’s Nano Banana Pro and Microsoft’s MAI-Image-2.

ChatGPT Images 2.0 is available to all ChatGPT and Codex users starting today.

IBL News | New York

Google CEO Sundar Pichai announced yesterday at the opening of the Google Cloud Next conference in Las Vegas, the Gemini Enterprise Agent Platform.

Google’s platform, delivered through the Gemini Enterprise app, is geared toward IT and technical teams and is intended for building and managing agents at scale. It’s the company’s answer to Amazon’s Bedrock AgentCore and to Microsoft Foundry.

“It brings together the best of Vertex AI with transformational new features, including Agent Studio, Agent-to-Agent Orchestration, Agent Registry, Agent Identity, Agent Gateway, Agent Observability, and much more,” explained Thomas Kurian, CEO of Google Cloud.

The Agent platform performs trigger-based processes, edits files without switching apps, has an Inbox for managing agent activity, and offers Skills to create shortcuts for repetitive tasks and Canvas to create and edit files.

The underlying models are Google’s Gemini LLM, Nano Banana 2 image generator, and Anthropic’s Claude — with support for Claude Opus, Sonnet, and Haiku, including the new Opus 4.7 launched last week.

At the same event, Alphabet Inc.’s Google Cloud unveiled its eighth generation of TPUs, including the TPU 8t for training and the TPU 8i for inference, which will be generally available later this year. These tensor processing units, or TPUs, are a homegrown chip that’s designed to take on NVIDIA and become a greater force in AI

In addition, Google shared plans to turn the Chrome browser into an AI coworker for enterprise users at the workplace. It then uses AI to handle various tasks such as booking travel, entering data, scheduling meetings, and other related tasks in web-based work.

Google suggested that infusing AI into Chrome could be used to input information into the CRM based on content in a Google Doc, compare vendor pricing across tabs, summarize a candidate’s portfolio before an interview, pull key data from a competitor’s product page, and more.

These workflows will still require the physical user to manually review and confirm the AI’s input before any action.

The idea is to help speed up these more tedious tasks, freeing people to focus on what Google calls “strategic work.”