Small and energy-efficient models have garnered growing attention in recent months.
Our daily email brings you smart and engaging news and analysis on the biggest stories in business and finance. For free.
Microsoft’s patent for a language model that talks back signals Big Tech’s interest in multimodal AI.
A high-powered chatbot may not deliver the illusive dream of developing artificial intelligence that can teach itself.
As one legal expert put the issue: “The technology is moving faster than the law ever will.”
Microsoft hasn’t signed off on OpenAI’s dramatic reversal of its onetime plan to become a for-profit venture.
Alibaba wants to shave down its neural networks to make AI easier to deploy.
The US is trying to box China out of AI development. The problem is the Asian nation is also the source of invaluable AI talent.
While this is a recognized problem in face-reading AI models, Nvidia’s tech relies on synthetic data to achieve balance, which comes with caveats.
Quarterly earnings at tech giants Meta and Microsoft surged, indicating that multi-billion dollar AI investments are starting to pay off.
Microsoft wants its language models to be a little more adaptable.
The company’s latest filing adds another solution to several aiming to tackle the ever-present issue of bias at the source.
A recent IBM patent shows that it may be working on Grammarly for software developers.
IBM is booting up its domestic production, setting aside $150 billion to make computers in the US over the next five years.
Along with mitigating hallucinations, this tech creates an audit trail for more transparency between the model and its users.
One point Chinese AI companies including Tencent and DeepSeek emphasize about their new models: efficiency.
This new server farm announcement comes just after Apple CEO Tim Cook reportedly paid President Trump a visit.