Author HubMatthias Bastian

Matthias is the co-founder and publisher of THE DECODER, exploring how AI is fundamentally changing the relationship between humans and computers.
AI research
Update

So-called reasoning models are more efficient but not more capable than regular LLMs, study finds

AI in practice

Google leans on token metrics, not revenue, adding to bubble talk about AI growth

AI and society

The long-predicted deepfake dystopia has arrived with Sora 2

Ad
AI in practice
Short

Rumors on LinkedIn claim that ChatGPT is no longer allowed to give medical or legal advice, but OpenAI says that’s false. The company says the model’s behavior has not changed. Karan Singhal, OpenAI’s Head of Medical AI, says ChatGPT was never meant to replace expert advice, but can still help users understand complex medical or legal topics.

Screenshot via X

OpenAI’s usage policy change logs show no recent changes to how sensitive topics are handled. The most recent update on October 29, 2025, was made to "reflect a universal set of policies across OpenAI products and services."

via waybackmachine

OpenAI’s usage policy change logs show no recent changes to how sensitive content is handled. The latest update on October 29, 2025, was made to unify the rules across all products. A line warning about giving advice that “requires a license” was already in earlier versions. Older policies included similar notes, just without the licensing reference.

AI in practice
Short

OpenAI has signed a $38 billion multi-year deal with Amazon Web Services (AWS) to run and expand its AI models using AWS infrastructure. The partnership includes access to AWS UltraServers powered by hundreds of thousands of NVIDIA GPUs and scalable CPUs. The agreement runs through at least 2026, with extension options. OpenAI's flagship models, such as GPT-5, will remain exclusive to Microsoft Azure and OpenAI's own platform, except for its open-source models.

via X

The AWS deal adds to a string of recent partnerships by OpenAI: with Nvidia and Broadcom for at least 10 gigawatts of compute each, AMD for up to 6 gigawatts, and Oracle for 4.5 gigawatts.

Ad
Ad
Ad
Ad
AI in practice
Short

OpenAI is piloting Aardvark, a security tool built on GPT-5 that scans software code for vulnerabilities. The system is designed to work like a security analyst: it reviews code repositories, flags potential risks, tests whether vulnerabilities can be exploited in a sandbox, and suggests fixes.

In internal tests, OpenAI says Aardvark found 92 percent of known and intentionally added vulnerabilities. The tool has also been used on open source projects, where it identified several issues that later received CVE (Common Vulnerabilities and Exposures) numbers.

Aardvark's workflow: GPT-5 scans code, tests for vulnerabilities, and suggests fixes. | Image: OpenAI

Aardvark is already in use on some internal systems and with selected partners. For now, it's available only in a closed beta, and developers can apply here. Anthropic offers a similar open source tool for its Claude model.

Ad
Ad
AI research
Short

Microsoft and OpenAI have decided they'll define for themselves what counts as artificial general intelligence (AGI) and when they've supposedly achieved it. The two companies say they'll appoint a panel of experts to make that call, but they haven't said who will be on it, what criteria they'll use, or even what AGI actually means. In a joint podcast, Sam Altman and Satya Nadella made it clear there's no shared definition or timeline, not even between them.

AGI was once considered a scientific milestone: an AI system that can think, learn, and solve problems like a human. Now, it's become a bargaining chip in a contract between two tech giants. That shift turns AGI into a label that can be applied or withdrawn whenever it's convenient, stripping it of any real, objective meaning. Maybe that was the goal all along—hype up the AGI label, then let it fade away when it no longer serves their interests.

Google News

Bank details

IBAN: DE88 2507 0070 0053 0014 00
BIC: DEUTDE2HXXX
Account holder: Deep Content GmbH
Purpose: Support THE DECODER
AI in practice

OpenAI's new ChatGPT image model matches Google's Nano Banana Pro on complex prompts

AI research

More AI agents isn't always better, new Google and MIT study finds

AI in practice

GPT-5.2 lands to top Google's Gemini 3 in the AI benchmark game just four weeks after GPT-5.1

Google News