the-decoder7d
Deepseek's V3 is the latest example of state-controlled censorship in Chinese LLMs
Users have discovered a way to bypass Deepseek V3's content filters through prompt engineering. By asking the model to insert periods between letters, they can get it to provide more balanced or China ...
the-decoder6d
OpenAI's o1-preview model manipulates game files to force a win against Stockfish in chess
Instead of playing a proper game of chess against Stockfish, a dedicated chess engine, o1-preview figured out how to hack its test environment to force a win. According to Palisade Research, an AI ...
the-decoder4d
Microsoft's new Large Action Model can perform some tasks in Word
Microsoft researchers have developed what they're calling a "Large Action Model" (LAM) - an AI that can operate Windows programs on its own. This marks a shift from AIs that just talk about doing ...
the-decoder6d
OpenAI co-founder says new AI safety approach "may apply to AGI and beyond"
Instead of just learning from examples of good and bad behavior, their latest o-series models can understand and actively reason through specific safety guidelines, the company says. In one example ...
the-decoder5d
Nvidia snaps up Israeli AI infrastructure company Run:ai
Nvidia has finalized its acquisition of Run:ai, an Israeli startup specializing in AI infrastructure management software. In an interesting twist, the company plans to open-source its previously ...
the-decoder5d
Hugging Face's Smolagents framework simplifies building AI agents with just a few lines of code
Hugging Face just released "smolagents," a stripped-down open-source library that lets developers create AI agents with minimal code. The library itself is remarkably lightweight, containing just ...
the-decoder7d
Meta envisions social networks where AI characters coexist alongside human accounts
Meta is laying the groundwork for a future where AI-generated profiles will interact alongside regular users on Facebook and Instagram, with some of these tools already rolling out. According to ...
the-decoder11d
New AI system recognizes soccer fouls, evaluates severity, and provides commentary on key plays
A new AI system can now watch soccer games, identify key plays, and even provide commentary similar to that of a human announcer. The technology comes from researchers at Shanghai Jiao Tong University ...
the-decoder9d
Deepseek's $5.6M Chinese LLM wonder shakes up the AI elite
A Chinese startup is proving you don't need deep pockets to build world-class AI. Deepseek's latest language model goes head-to-head with tech giants like Google and OpenAI - and they built it for a ...
the-decoder10d
Deepseek-V3 emerges as China's most powerful open-source language model to date
DeepSeek, a Chinese AI company, has introduced DeepSeek-V3, its most powerful language model to date, featuring 671 billion parameters in a mixture-of-experts architecture. DeepSeek-V3 was trained on ...
the-decoder12d
OpenAI's o1-preview outperforms doctors in diagnosing tricky medical cases, study finds
A new study suggests OpenAI's o1-preview AI system might be better at diagnosing tricky medical cases than human doctors. A team of researchers from Harvard Medical School and Stanford University put ...
the-decoder7d
Google currently has no plans for a test-time premium subscription similar to ChatGPT Pro
Google's leadership is gearing up for what they see as a pivotal year in artificial intelligence. At a recent internal strategy meeting, CEO Sundar Pichai didn't mince words about the company's ...