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

OpenAI is rolling out "ChatGPT for Teachers," a free version of its AI chatbot for verified K-12 teachers in the United States. The offer runs through June 2027 and includes a secure workspace that, according to the company, does not use data for model training by default. OpenAI says teachers are already seeing time savings in lesson planning and other daily tasks.

"Every student today is growing up with AI, and teachers play a central role in helping them learn how to use these tools responsibly and effectively."

OpenAI

The platform meets US privacy standards like FERPA and gives teachers access to the GPT-5.1 Auto model along with integrations for Canva and Google Drive. School administrators can manage and assign licenses centrally. OpenAI is also partnering with groups like the American Federation of Teachers to help educators learn how to use the technology effectively.

 

Ad
Ad
Ad
Ad
AI in practice
Short

Google is updating the Gemini app with a new way to control its AI video model. With the latest release, users can upload multiple reference images for a single video prompt. The system then generates video and audio based on those images combined with text, giving people more direct control over how the final clip looks and sounds.

Google previously tested this feature in Flow, the company's expanded video AI platform. Flow also supports extending existing clips and stitching together multiple scenes, and it offers a slightly higher video quota than the Gemini app. Veo 3.1 has been available since mid-October and, according to Google, delivers more realistic textures, higher input fidelity, and better audio quality than Veo 3.0.

Ad
Ad
AI and society
Short

Yann LeCun accuses Anthropic of regulatory capture. The dispute centers on an AI-driven cyberattack that Anthropic says happened with almost no human oversight and posed a serious cybersecurity threat. After the company published its findings, US Senator Chris Murphy called for tougher AI regulation.

Chris Murphy and Yann LeCun reacted publicly after Anthropic warned about a large-scale AI-driven cyberattack. | Image: X

LeCun, who reportedly is preparing to leave Meta, pushed back on the political reaction and accused companies like Anthropic of using questionable studies to stoke fear and push for stricter rules that would disadvantage open models. In his view, the goal is to shut out open-source competitors.

Trump's AI advisor, David Sacks, has also accused Anthropic of using what he called a "sophisticated regulatory capture strategy based on fear-mongering."

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

DeepseekMath-V2 is Deepseek's latest attempt to pop the US AI bubble

AI research

Frustrated authors withdraw papers after realizing their reviewers are just lazy LLMs

AI research

Gemini 3 Pro tops new AI reliability benchmark, but hallucination rates remain high

Google News