Hub Artificial Intelligence

Artificial intelligence is either a field of research, the most important technical innovation of mankind, its downfall – or simply a pipe dream from Silicon Valley. Can machines really be intelligent? What is intelligence anyway? What opportunities does AI technology offer, and what are the risks? From neural networks to the science fiction vision of super AI, from deepfakes to AI surveillance: THE DECODER delivers the latest AI news and information on all facets of artificial intelligence.

Artificial Intelligence

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

Artificial Intelligence

Researchers push "Context Engineering 2.0" as the road to lifelong AI memory

Artificial Intelligence
Update

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

Ad
Artificial Intelligence
Short

According to a study in Nature, Chinese AI company Deepseek trained its R1 language model for only $294,000. The run used 512 Nvidia H800 chips developed specifically for the Chinese market. Nvidia confirmed the chips were delivered before U.S. export restrictions took effect. In its technical paper, Deepseek also admitted to using A100 GPUs during preparation for a smaller prototype, after U.S. officials had earlier suspected the company of holding unauthorized H100s.

The figure does not include the much larger costs of training Deepseek’s underlying V3 foundation model. Estimates for that project vary widely, ranging from tens of millions to several hundred million dollars depending on the source.

Deepseek’s claim of unusually low training costs previously rattled global tech markets, triggering sharp declines in the share prices of major AI hardware and software companies.

Artificial Intelligence
Short

Meta is in negotiations with publishers including Axel Springer, Fox Corp., and News Corp. about licensing deals that would allow the company to use news articles in its AI products, such as chatbots. The talks were first reported by the Wall Street Journal.

The company has already secured a licensing agreement with Reuters, but its broader discussions with media companies have only recently begun. For Meta, the move marks a shift: in 2022 it shut down Facebook’s News Tab and stepped back from directly funding journalism.

Other tech companies have moved earlier in this area. OpenAI has inked content deals with publishers such as Hearst, while Amazon has also pursued licensing agreements. Rising resistance from publishers over the use of their material for AI training has become the backdrop for these negotiations.

WSJ
Ad
Ad
Ad
Ad
Ad
Ad
Artificial Intelligence
Short

Google has introduced the Agent Payments Protocol (AP2), a new open standard aimed at enabling AI agents to carry out secure payments across different platforms. AP2 builds on the existing Agent2Agent protocol and supports a wide range of payment methods, including credit cards, stablecoins, and bank transfers.

A key element of the design is digital mandates, which are cryptographic authorizations that lock in user intent. These mandates are intended to ensure that transactions remain verifiable and secure, whether for real-time purchases or automated transactions when the user is not directly involved.

The initiative is already backed by more than 60 companies, among them Mastercard, PayPal, Coinbase, and Adobe. By creating a unified framework, the goal is to establish a standardized and trustworthy system for agent-driven commerce. Google has made the documentation available on GitHub.

Google News

Bank details

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

AI research

Corporate AI agents use simple workflows with human oversight instead of chasing full autonomy

Google News