Artificial Intelligence

Mar 7, 2026

OpenAI is launching a new support program for open-source developers. Core maintainers of public software projects can apply for six months of free access to ChatGPT Pro with Codex, API credits, and Codex Security. Codex Security, a new AI tool for code security checks, will be reviewed on a case-by-case basis and only granted selectively due to the capabilities of GPT-5.4, according to OpenAI.

Developers who prefer other programming tools like OpenCode, Cline, or OpenClaw can also apply. Projects that don't meet all the criteria but play an important role in the broader software ecosystem are also welcome to apply. The program builds on OpenAI's existing Codex Open Source Fund, which the company has backed with one million dollars.

Comment Source: OpenAI

Matthias Bastian

Mar 7, 2026

AI in practice

Anthropic's Claude Code subscription may consume up to $5,000 in compute per month while charging the user just $200

Anthropic’s $200 Claude Code subscription could consume up to $5,000 in compute per user, according to Cursor’s internal analysis reported by Forbes. The numbers reveal just how aggressively AI companies are subsidizing their coding tools and what that could mean for prices once these tools become essential.

Read full article

Comment

Jonathan Kemper

Mar 7, 2026

AI research

Bytedance's open-weight Helios model brings minute-long AI video generation close to real time

Yellow sports car on forest road with overlay text about AI video generation by Helios

Matthias Bastian

Mar 7, 2026

Short News

OpenAI and Oracle have decided not to expand their data center site in Abilene, Texas, beyond the planned 1.2 gigawatts. Oracle has leased eight buildings at the location for OpenAI, designed to house around 400,000 Nvidia Blackwell chips, but only two have been completed so far.

Oracle had pushed to get OpenAI into six more buildings, but both sides passed because the additional power supply wouldn't be available for at least a year. Instead of expanding the current Blackwell generation, OpenAI plans to buy Nvidia's next-generation Vera Rubin chips for a different data center. According to Bloomberg, Nvidia is now trying to get Meta to fill the vacant space, though those talks are still in the early stages.

OpenAI's compute manager Sachin Katti described the Stargate site as already one of the largest AI data center campuses in the country. "We considered expanding it further, but ultimately chose to put that additional capacity in other locations," Katti writes, adding that OpenAI is currently developing more than half a dozen sites across several US states.

Comment Source: Bloomberg

Matthias Bastian

Mar 7, 2026

Short News

Anthropic's coding tool Claude Code now supports local, scheduled tasks through a new /loop command. Users can set up recurring jobs at fixed intervals—minutes, hours, or days—that run in the background as long as Claude Code is active and auto-delete after three days. The feature uses standard cron expressions and the local time zone. One-time natural language reminders like "remind me at 3 PM to push the release branch" are also supported, with up to 50 scheduled tasks per session.

Anthropic developer Thariq Shihipar gives the example of checking error logs every few hours, with Claude Code automatically creating pull requests for fixable bugs. The feature gets especially interesting when connected to other data sources, he says. Claude Code creator Boris Cherny adds use cases like auto-monitoring pull requests with self-fixing or generating morning Slack summaries. A detailed guide is available here.

Claude Code recently received updates adding automated desktop functions, remote control for smartphones, and memory.

Comment Source: via X

Matthias Bastian

Mar 7, 2026

Short News

Google, Amazon, and Microsoft are assuring customers that Anthropic's AI technology will remain available outside of defense projects, CNBC reports, citing company spokespeople. None of the three have issued public statements.

Microsoft said its legal team reviewed the classification and concluded that Anthropic products, including Claude, can stay available through M365, GitHub, and Microsoft's AI Foundry, except for the US Department of War. Google continues offering Claude through Vertex AI, Amazon through Bedrock and GovCloud.

The reassurances come amid an ongoing dispute between Anthropic and the US Department of War, which designated Anthropic as a supply chain risk. CEO Dario Amodei called the designation legally untenable and said his company plans to fight it in court. Some defense contractors have already told employees to switch to alternatives like OpenAI.

According to the Washington Post, Claude was still used in the recent US attack on Iran despite the designation. The transition to other AI models is expected to take six months, with OpenAI among the new partners.

Comment Source: CNBC

Matthias Bastian

Mar 7, 2026

Short News

Anthropic is launching the Anthropic Marketplace, a storefront where corporate customers can buy third-party software built on Anthropic's AI models. Launch partners include Snowflake, Harvey, and Replit. Anthropic isn't taking a commission, and customers can put a portion of their existing annual Anthropic spend toward these tools.

First providers on the Anthropic Marketplace. | Image: Screenshot

The strategy mirrors what Amazon does with the AWS Marketplace or Microsoft with Azure: Anthropic is positioning itself as the orchestrator of a growing AI software ecosystem. By waiving commissions and letting customers tap existing budgets for third-party tools, the company makes it easy to stay in its ecosystem.

The commission waiver is less generous than it sounds, though. Every marketplace app runs on Anthropic's models, so partners already pay for the model capacity they use. Anthropic doesn't earn from the marketplace itself (yet), but it profits from the growing demand for its models that the marketplace generates. Interested partners can sign up for the waiting list.

Comment Source: Anthropic

Jonathan Kemper

Mar 7, 2026

Update

AI research

Video AI models hit a reasoning ceiling that more training data alone won't fix, researchers say

Filmstrip collage: Video reasoning tasks such as maze navigation, 3D rotation, tile puzzles, object counting and physical predictions.

Matthias Bastian

Mar 6, 2026

Short News

OpenAI launches "Codex Security," an AI-powered security agent built to find vulnerabilities in software projects. The tool, formerly known as "Aardvark," is now available as a research preview for ChatGPT Enterprise, Business, and Edu customers, free for the first month. Codex Security analyzes code repositories, builds a project-specific threat model, and tests any vulnerabilities it finds in isolated test environments, OpenAI says.

During the beta phase, OpenAI says the system cut false positives by more than 50 percent and reduced redundant alerts by 84 percent in one case. Over the past 30 days, Codex Security scanned more than 1.2 million commits and flagged 792 critical vulnerabilities.

OpenAI has also reported vulnerabilities in open-source projects, including OpenSSH, GnuTLS, GOGS, Thorium, and Chromium, with 14 CVEs issued so far. A program for open-source maintainers is set to expand. More details on getting started are available in the documentation.

Anthropic also recently shipped its own cybersecurity tool, sending cybersecurity stocks into the red.

Comment Source: OpenAI

Maximilian Schreiner

Mar 6, 2026

Short News

The Pentagon formally notified Anthropic on March 4 that the company and its products have been designated as a supply chain risk to US national security. Anthropic CEO Dario Amodei announced on March 5 that the company will challenge the designation in court. According to Amodei, the measure only applies to the use of Claude as a direct part of contracts with the Department of War, not all customers.

The dispute escalated after Anthropic demanded assurances that its AI would not be used for mass surveillance of Americans or autonomous weapons deployment. According to Bloomberg, Claude is currently the only AI system operating in the Pentagon's classified cloud and is actively used in US operations against Iran.

Amodei also apologized for an internal post that was leaked to the press. "It does not reflect my careful or considered views," Amodei said. The post was written within hours of Trump's Truth Social post, Secretary of War Hegseth's announcement on X, and a Pentagon deal with OpenAI.

Anthropic is offering to continue providing its models to the military at nominal cost until a transition is complete. According to Bloomberg, the company is currently valued at $380 billion and is approaching an annual revenue run rate of $20 billion.

Comment Source: Anthropic