DeepLearning.AI - The Batch | DeepLearning.AI (Page 3)

Team in modern office applauding while watching a news anchor on a big screen.

Data Points

All the models we’ve been waiting for: OpenAI’s scaled-up Project Orion arrives

Mercury debuts diffusion language models. Alibaba’s top video model is now free to download. A new model from Tencent is built for speed. IBM’s Granite 3.2 models are built for business.

Diagram comparing direct audio generation with a foundation model vs. a voice pipeline using STT, LLM, and TTS.

Letters

What I’ve Learned Building Voice Applications: Best practices for building apps based on AI’s evolving voice-in, voice-out stack

The Voice Stack is improving rapidly. Systems that interact with users via speaking and listening will drive many new applications.

The Batch Newsletter

Meta Reads Minds, Big AI Spending Climbs, Deepfakes Appropriate Celeb Likenesses, Reasoning in Vectors

The Batch AI News and Insights: The Voice Stack is improving rapidly. Systems that interact with users via speaking and listening will drive many new applications.

Diagram of Coconut, a method training LLMs to process thought chains as vectors, comparing it to Chain-of-Thought (CoT).

Machine Learning Research

Reasoning in Vectors, Not Text: Meta introduces Chain of Continuous Thought (Coconut) to improve next-token prediction

Although large language models can improve their performance by generating a chain of thought (CoT) — intermediate text tokens that break down the process of responding to a prompt into a series of steps.

“Enough Is Enough” in black on white, with a heart, from an AI-generated video protesting Kanye West’s antisemitism.

Tech & Society

Deepfake Developers Appropriate Celebrity Likenesses: Viral video uses AI to depict celebrities without consent, sparking legal debate

A viral deepfake video showed media superstars who appeared to support a cause — but it was made without their participation or permission.

Bar chart of 2024 vs. 2025 capital expenditures for Amazon, Microsoft, Alphabet, and Meta, showing projected increases.

Business

Big AI Spending Continues to Rise: Tech giants increase cloud spending to meet growing infrastructure demands

Top AI companies announced plans to dramatically ramp up their spending on AI infrastructure.

A participant types while an MEG scan decodes brain activity into text in real-time, showing typed vs. decoded text.

Science

Reading Minds, No Brain Implant Required: Brain2Qwerty, a system that decodes thoughts using brain waves without surgery

To date, efforts to decode what people are thinking from their brain waves often relied on electrodes implanted in the cortex. New work used devices outside the head to pick up brain signals that enabled an AI system, as a subject typed, to accurately guess what they were typing.

A minimalist high-tech laboratory with two scientists working.

Data Points

Anthropic releases Claude 3.7 Sonnet as a hybrid reasoning model: DeepSeek’s FlashMLA is its first entry in OpenInfra week

Figure’s Helix vision language action robotics model. Google fine-tunes its own family of open VL models. SuperGPQA may be the most challenging general knowledge test yet. Meta creates new framework to evaluate agentic LLMs.

A realistic depiction of a high-tech research laboratory where robotic assistants aid scientists in conducting experiments.

Data Points

Microsoft builds a generative world-action model: Perplexity opens up uncensored “1776” version of DeepSeek-R1

Researchers develop a highly capable language diffusion model. Google’s hypothesis-making agent is your new research partner. A new family of vision-language models for low-resource devices. HP will brick Humane’s Ai Pin and repurpose its tech for new devices.

Thermal aerial image showing a suspect surrendering with hands raised. A marker highlights their location.

Letters

How AI Saved a Police Officer’s Life: A police drone located an officer under attack before it was too late. Get ready for many more stories like this.

Last month, a drone from Skyfire AI was credited with saving a police officer’s life after a dramatic 2 a.m. traffic stop.

The Batch Newsletter

Grok 3 Scales Up, Mobile Apps Generated To Order, Musk Moves On OpenAI, Officials Reverse Course on AI Regulation

The Batch AI News and Insights: Last month, a drone from Skyfire AI was credited with saving a police officer’s life after a dramatic 2 a.m. traffic stop.

AI Action Summit in a grand hall with domes, flags, and a crowd attending the event.

Tech & Society

World Powers Move to Lighten AI Regulation: Global AI summit reveals deep divisions on regulation and governance

The latest international AI summit exposed deep divisions between major world powers regarding AI regulations.