Releases

Google Cloud launches Vertex AI as a unified platform for building, deploying and operating ML models.

Google introduces LaMDA as a conversational language model designed for open-ended dialogue.

Hugging Face introduced a paid individual plan with expanded private, storage, and compute capabilities on the Hub.

Hugging Face helped launch the BigScience open research workshop, coordinating an international collaboration to train a large multilingual language model.

Hugging Face launched Accelerate to let PyTorch code run across distributed hardware with minimal changes.

Scale raised $325 million in Series E funding co-led by Dragoneer, Greenoaks, and Tiger Global at a reported $7.3 billion valuation.

Microsoft agrees to acquire Nuance Communications for $19.7 billion including net debt, expanding its healthcare and conversational-AI capabilities.

NVIDIA introduced Morpheus for real-time AI-assisted threat detection using GPUs and BlueField DPUs.

NVIDIA introduced Base Command to provision, schedule and operate shared DGX SuperPOD AI infrastructure.

Baidu's Class A ordinary shares began trading on the Hong Kong Stock Exchange under 9888.

Hugging Face raised $40 million in Series B funding led by Addition to grow its open-source machine-learning ecosystem and hosted platform.

NVIDIA announced an enterprise software suite for virtualized AI workloads on VMware vSphere.

Databricks and Google Cloud launched Databricks on Google Cloud, completing availability across the three major clouds.

Databricks raised $1 billion led by Franklin Templeton at a $28 billion post-money valuation.

OpenAI introduced CLIP, a neural network trained on image-text pairs for transferable visual recognition.

OpenAI introduced DALL-E, a model that generates images from text descriptions.

Erik Bernhardsson and Akshat Bubna founded Modal to build a developer-friendly serverless cloud for compute-intensive workloads.

Anthropic was founded at the beginning of 2021 as an AI safety and research company led by Dario and Daniela Amodei.

DeepMind publishes MuZero, which plans using a learned model without being told the environment's underlying rules.

NVIDIA opened the Omniverse collaboration and simulation platform to beta users.

AWS introduced a HIPAA-eligible health data lake using FHIR and machine learning.

AWS introduced tools for bias detection and model explainability.

Scale acquired real-time video-understanding startup Helia AI and brought its team and infrastructure expertise into the Nucleus platform.

Scale raised $155 million in a Series D led by Tiger Global at a post-money valuation above $3.5 billion.

AWS introduced machine-learning-based anomaly detection for business and operational metrics.

AWS introduced an end-to-end condition-monitoring system using sensors and machine learning.

AWS introduced predictive maintenance from industrial sensor data.

AWS introduced a computer-vision service for visual defect and anomaly detection.

AWS introduced an ML-powered service for detecting abnormal operational behavior and recommending remediation.

AWS introduced pretrained models and solution templates to accelerate machine-learning deployment.

AWS announced its first custom chip purpose-built for high-performance machine-learning training.

AlphaFold 2 demonstrates accuracy competitive with experimental structures at CASP14, marking a major protein-folding breakthrough.

Transformers v4 began its release cycle with broad framework and architecture support and a standardized model API.

NVIDIA introduced a deskside AI system with four A100 GPUs for workgroups.

NVIDIA doubled A100 accelerator memory to 80GB using HBM2e.

Facebook AI released M2M-100, a multilingual machine-translation model translating directly between 100 languages.

Baidu opened Apollo Go autonomous ride-hailing to public users in Beijing.

NVIDIA introduced Maxine, a cloud-native suite of GPU-accelerated AI video-conferencing and streaming software.

Scale released the core Nucleus workspace for dataset visualization, search, curation, collaboration, and model-error analysis.

Scale launched Nucleus as a dataset-management and model-debugging workspace for visualizing, curating, searching, and collaborating on machine-learning data.

Scale introduced a self-serve, pay-as-you-go product for receiving production-quality labels and instruction feedback in as little as one hour.

Microsoft spins Xiaoice into an independent company while retaining an investment interest.

Scale launched a mapping product for computer-vision and autonomous path-planning workflows.

Hugging Face began its hosted inference platform with a serverless API for models on the Hub, later expanding into dedicated endpoints and provider routing.

Hugging Face introduced a hosted API for running models from the Hub without managing serving infrastructure.

Databricks acquired open-source visualization company Redash to expand SQL analytics and collaboration.

Facebook AI introduced wav2vec 2.0 for self-supervised speech learning with limited labeled data.

OpenAI launched a general-purpose API for accessing its language models.

Microsoft Research introduces DeBERTa, improving Transformer language models with disentangled attention and enhanced mask decoding.

OpenAI published GPT-3, demonstrating few-shot learning at 175 billion parameters.