16 posts
GitHub vai encerrar o GitHub Models em 30 de julho, com brownouts antes da data. O fim do playground, da API e do BYOK força times a rever stack, governança e custo.
GitHub is retiring GitHub Models on July 30, with brownouts beforehand. Losing the playground, API, and BYOK forces teams to rethink stack, governance, and cost.
O Google reposiciona o Spanner como base relacional, vetorial, grafo e full-text para agentes. A promessa é trocar arquitetura fragmentada por contexto operacional unificado.
Google is repositioning Spanner as a unified relational, vector, graph, and full-text foundation for agents. The pitch is less glue, more operational context.
Claude Sonnet 5 arrives in GitHub Copilot with lower launch pricing, API migration changes, and direct implications for coding-agent cost and enterprise governance.
Claude Sonnet 5 estreia no GitHub Copilot com preço menor, mudanças de migração na API e impacto direto no custo e na governança dos agentes de código.
NVIDIA is bundling Agent Toolkit, OpenShell, NemoClaw and Nemotron into a safer, more auditable operating stack for enterprise AI agents.
AWS is turning live web retrieval into a governed AgentCore capability, not just a search API hookup, with zero data egress and MCP-native delivery.
Benchmark MosaicLeaks mostra como agentes de pesquisa que cruzam dados internos com busca na web podem vazar informação sensível sem perceber.
MosaicLeaks shows how research agents can expose private data through normal-looking web searches, and why privacy-aware training matters.
OpenAI’s Ona deal points to trusted cloud workspaces where agents can run persistently, securely, and under customer control.
Claude Design matters less as a flashy demo and more as Anthropic’s attempt to turn Claude into a governed, collaborative creative workflow.
GPT-5.5, GPT-5.4 e Codex no Amazon Bedrock reduzem atrito de compra, segurança, compliance e deploy para empresas que já operam em AWS.
OpenAI models and Codex on AWS matter because they fit existing enterprise procurement, security, and deployment workflows, not just because they add another channel.
ITBench-AA shows frontier models score below 50% on real SRE tasks. Here's why this matters for enterprise agent adoption.
ITBench-AA shows frontier models score below 50% on real SRE tasks. Here's why this matters for enterprise agent adoption and what it means for your AI strategy.