Author: IBL News

IBL News | New York

Anthropic, the owner of the chatbot Claude, made its latest coding model, Sonnet 4.5, available this week. Developers can use the Claude API.

The tool includes checkpoints, a feature that allows users to save progress and roll back instantly to a previous state. The company reports substantial gains in reasoning and math.

Anthropic explained:

• “We’ve refreshed the terminal interface and shipped a native VS Code extension. We’ve added a new context editing feature and memory tool to the Claude API that lets agents run even longer and handle even greater complexity.”

• “In the Claude apps, we’ve brought code execution and file creation (spreadsheets, slides, and documents) directly into the conversation.”

• “We have made the Claude for Chrome extension available to Max users who joined the waitlist last month.”

• “We’re also giving developers the building blocks we use ourselves to make Claude Code. We’re calling this the Claude Agent SDK. The infrastructure that powers our frontier products—and allows them to reach their full potential—is now yours to build with.”

IBL News | New York

OpenAI released its latest video and audio generation model, Sora 2, this week, aiming to establish a competitive position against more advanced offerings, such as Google’s Veo 3, Runway, and KlingAI.

This new AI video generator is available in the U.S. and Canada as a free iOS app, simply called “Sora”.

ChatGPT Pro users will also be able to use the experimental, higher-quality Sora 2 Pro model on sora.com⁠ (and soon in the Sora app as well).

The new model also allows the injection of elements from the real world, such as cameos, enabling people to insert themselves into any AI-generated video scene.

The feature of “cameos” can generate significant social traction. “Overwhelming feedback from testers is that cameos are what make this feel different and fun to use, as it is a new and unique way to communicate with people.” 

OpenAI is rolling this out as an invite-based app “to make sure you come in with your friends.”

Only users can decide who can use their own cameos, and they can revoke access or remove any video that includes it at any time. Videos containing cameos, including drafts created by other people, are viewable by the users at any time.

OpenAI has announced that it is implementing default limits on the number of generations teens can view per day in the feed, and it’s also rolling out stricter permissions on Cameos for this group. In addition, it’s scaling up teams of human moderators to quickly review potential cases of bullying.

Sora 2 creates background soundscapes, speech, and realistic sound effects. along with realistic and cinematic scenes.

Though it still makes plenty of mistakes, Sora 2 is better about obeying the laws of physics compared to prior systems. “It excels at realistic, cinematic, and anime styles,” said the company.

The San Francisco-based research lab compared Sora 2 to GPT-3.5, while its original, overoptimistic model of Sora would be GPT-1.

OpenAI plans to release Sora 2 as an API, while Sora 1 Turbo will remain available, and all content created will continue to reside in the existing sora.com library.