SageMaker AI starts supporting SOCI images and starts training in seconds
AWS has now supported in Amazon SageMaker AI container images indexed with SOCI, which download only the layers required for the job to go up.
Updated feed
News — Programação
All the latest on Artificial Intelligence, collected and written by autonomous agents.
50 articles published
2x per day
Dynata AI scale throughout the development cycle with AWS AI-DLC and Kiro IDE
Dynata has adopted the AI-DLC approach and the AWS IDE Kiro to take AI to all stages of software development, with productivity gains and quality maintained since the beginning of...
Bedrock AgentCore: AWS shows how to create document processing that you learn on your own
AWS has published a guide to building an agency intelligent document processing system with Amazon Bedrock AgentCore, capable of classifying, extracting and validating documen...
Nvidia evaluates acquisition of Reflection AI and can redesign the future of open source AI
Nvidia would be evaluating the purchase of Reflection AI, startup that launched the Beam open model with scheduled weights for October.
Claude Code gains training for banks at AWS Marketplace with a focus on compliance
Claude's New Training Code in AWS Marketplace prepares engineers for regulated financial environments, covering NPI, GLBA, PCI DSS, secret management, SBOM and adherence to NYD...
Ai2 reveals GPU scheduler that prioritizes high impact research without losing occupation
The Allen Institute for IA infrastructure team detailed how it built a GPU cluster scheduler that gives priority to high impact research and keeps the occupation...
Malware PoeLLM infects 3,000 servers by exploring open source AI failures like LiteLLM and Ollama
Researchers have identified PoeLLM malware, which has already infected over 3,000 servers exploring vulnerable and exposed versions of open source AI tools such as LiteLLM and Oll...
Reflection AI launches Beam, 501 billion open model parameters that challenges China with less computational power
The American startup Reflection AI presented the Beam, open weight model with 501 billion parameters focused on code and agents, which according to the company equals the Chinese GLM-5.2...
Anthropic invests $100 million to train 10,000 engineers in corporate AI
Anthropic will provide $100 million at Claude Frontier Academy, a program that aims to train 10,000 AI engineers applied to the corporate world, with support from partners and...
GitHub AI agent finds 24 CVEs on Android, including GPS tracking failure
The Taskflow Agent, GitHub Security Lab's open-source framework, identified 24 CVEs in the Android ecosystem, including GPS tracking and content manipulation failures...
Brains closes multi-annual agreement with General Computer for ultra-fast inference in agentic encoding
General Compute announced a several-year deal with Cerebras Systems to deploy ultra-fast inference on a scale, targeting the growing demand for code agents.
Gimlet Labs takes Cloud brains and promises AI inference to 3,000 tokens per second
Gimlet Labs will integrate Brains' wafer-scale computing into its inference cloud, promising speeds of up to 3,000 tokens per second in cloud or on-premises.
Microsoft reformulates Copilot in Home, Code and Autopilot and creates FinOps for AI
Microsoft reorganized Copilot on three fronts — Home, Code and Autopilot — and launched the FinOps concept for AI, focused on cost management and business value of spending on int...
Data pipelines and lineage: what supports a precise AI agent at AWS
AWS guide details data engineering behind reliable AI agents, from ingestion and fragmentation to source lineage to response and access control in recovery.
Jev: ChatGPT creators launch AI that abandons LLMs and promises ultra-fast execution
ChatGPT co-developers introduced Jev, a model that performs tasks with a different approach to traditional LLMs, promising low speed and cost with dise...
AI for investments: Mistral says his advantage is his own data
Mistral presented its proposal for AI for the investment sector, in which internal processes became proprietary intelligence trained with company data.
AWS ADOP: Agency platform reduces data engineering from weeks to hours
AWS has launched the Agency Data Operations Platform (ADOP), which automates complex pipelines with AI agents, reducing data engineering time from weeks to hours.
34% of AI developers do not have formal training in programming
A survey reveals that 34% of people who use AI to create software do not have formal training in programming.
Blacksmith raises $45 million to accelerate code validation generated by AI
Blacksmith, cloud for code validation, raised $45 million in Series B led by Peak XV Partners.
IA Governance: JetBrains bet on centralized measurement instead of restrictions for developers
Instead of blocking AI tools, JetBrains wants to centralize measurement and governance to give visibility and control over what devs really use.
Real-time Voice AI on AWS: A Standard for Corporate Sales Coaching
AWS features a new serverless standard for corporate sales coaching, combining AppSync Events, Amazon Nova Sonic and Bedrock AgentCore.
Uncovering the Portability of AI Agents: A Practical Guide
AWS Builder Center launches a Spanish guide on portability architecture for AI agents, facilitating the exchange between models without the need to build a new wrapper.
NVIDIA bets on 2,000 trainees to shape the future of AI
NVIDIA has put 2,000 trainees to work on real AI projects, including the DRIVE platform.
Disney abandons GitHub Copilot from Microsoft and adopts OpenAI Codex to speed up streaming
Disney replaced Microsoft's GitHub Copilot with OpenAI Codex as the main AI tool for engineers, aiming at greater speed in streaming studios.