OpenAI shares early data on coding agents in AI research
The company’s account focuses on agent usage, experiment velocity and task complexity, but the supplied evidence does not include the underlying findings.
Sep 6, 2026
Topics
A growing collection of articles on AI: big questions, useful tools, and ideas that stay with you.
Real ideas / Practical perspectives / A brighter nextThe company’s account focuses on agent usage, experiment velocity and task complexity, but the supplied evidence does not include the underlying findings.
Sep 6, 2026
A nightly AWS workflow combines TTL expiration, relevance scoring, LLM-based consolidation, regression testing, and deletion controls to keep long-running agents’ memories useful, auditable, and manageable.
Sep 6, 2026
The open-source control plane gives infrastructure teams a web interface, REST APIs and an AI agent backed by the same validation, reconciliation and persisted state for…
Sep 6, 2026
The release combines 786 pull requests from 214 contributors, adding nine model entries, beam search, DeepEP v2, broader speculative decoding, unified caching, diffusion improvements and extensive…
Sep 6, 2026
AWS outlines a persistent, shared GPU architecture for synthetic data generation, distributed post-training, and closed-loop evaluation, with end-to-end GPU goodput—not isolated job throughput—as the central operating…
Sep 6, 2026
AWS has opened access to GPT-5.6 Sol, Terra, and Luna through Amazon Bedrock endpoints in Sydney and Melbourne, with three API paths, prompt caching, federated Codex…
Sep 3, 2026
Jamf’s production architecture turns Bedrock invocation logs into daily per-user cost estimates, then uses scheduled Lambda processing and IAM Customer Managed Policies to apply tiered model…
Sep 3, 2026
The x402-secure system checks endpoints before agents pay them, while session spending limits, isolated credentials, role separation and audit logs constrain what happens when software transacts…
Sep 3, 2026