https://signal.yashgarg.me/A focused briefing on the latest artificial intelligence developments.en-ushttps://signal.yashgarg.me/articles/openai-gsa-onegov-government-access/https://signal.yashgarg.me/articles/openai-gsa-onegov-government-access/OpenAI and the U.S. General Services Administration have announced a 27-month OneGov agreement that removes platform fees and cuts consumption-based ChatGPT usage 50% for eligible federal, state, local and tribal governments starting October 1. The package also expands access and training for verified government cyber defenders.Fri, 11 Sep 2026 22:39:51 GMTGovernment AIhttps://signal.yashgarg.me/articles/openai-codex-chatgpt-antimicrobial-discovery/https://signal.yashgarg.me/articles/openai-codex-chatgpt-antimicrobial-discovery/OpenAI’s latest Applied AI case study follows César de la Fuente’s University of Pennsylvania lab as it combines its own deep-learning models with ChatGPT and Codex to search living and extinct genomes for potential antimicrobial molecules. The tools help the transdisciplinary team brainstorm, write code, process sequence data, and analyze results; experiments still determine whether any candidate works.Fri, 11 Sep 2026 21:30:45 GMTApplied AIhttps://signal.yashgarg.me/articles/openai-habitat-storage-one-billion-users/https://signal.yashgarg.me/articles/openai-habitat-storage-one-billion-users/OpenAI has detailed how Habitat, the online storage layer behind ChatGPT, the API, Codex, and internal services, grew from a small Python library into a distributed platform handling more than 70 million requests per second and more than 500 petabytes. The company says a two-engineer team used Codex and GPT-5.5 to rewrite 95% of the service in Rust.Fri, 11 Sep 2026 20:38:03 GMTAI infrastructurehttps://signal.yashgarg.me/articles/openai-agents-api-cloud-harness/https://signal.yashgarg.me/articles/openai-agents-api-cloud-harness/OpenAI has put the Agents API into public beta, giving developers a managed harness for long-running cloud agents, tool use, context management, and parallel subagents. Builders can choose OpenAI-hosted sandboxes, their own infrastructure, or a supported partner environment.Fri, 11 Sep 2026 19:31:43 GMTAI infrastructurehttps://signal.yashgarg.me/articles/openai-chatgpt-financial-services/https://signal.yashgarg.me/articles/openai-chatgpt-financial-services/OpenAI has introduced ChatGPT for Financial Services, a tailored ChatGPT Work experience that combines GPT-6 Astra with built-in financial datasets, citation-aware analysis, and enterprise controls. The initial design partnership with Morgan Stanley and Evercore focuses on investment banking and equity research workflows.Fri, 11 Sep 2026 19:31:43 GMTEnterprise AIhttps://signal.yashgarg.me/articles/openai-data-agent-chatgpt-work/https://signal.yashgarg.me/articles/openai-data-agent-chatgpt-work/OpenAI has launched a Data agent in ChatGPT Work that connects approved business sources, investigates changes in company metrics, and produces interactive dashboards through plain-language prompts. Administrators can enforce existing data permissions and let teams share findings or take approved actions through connected tools.Fri, 11 Sep 2026 19:31:43 GMTEnterprise AIhttps://signal.yashgarg.me/articles/openai-gpt-live-1-api-full-duplex-voice/https://signal.yashgarg.me/articles/openai-gpt-live-1-api-full-duplex-voice/OpenAI says GPT-Live-1 is now available in its API as a front-end voice layer that can listen and speak at the same time. The release targets lower-latency voice agents, delegates deeper reasoning and tool calls to backend models, and starts at $0.05 per minute.Fri, 11 Sep 2026 17:46:45 GMTAI productshttps://signal.yashgarg.me/articles/anthropic-tactical-intelligence-conventional-weapons-evaluations/https://signal.yashgarg.me/articles/anthropic-tactical-intelligence-conventional-weapons-evaluations/Anthropic’s Frontier Red Team reports new evaluations of location-targeting from fragmentary clues and engineering drones against moving targets. The company says frontier models are making consistent progress on simulated intelligence and weapons-development tasks, while open-weight models still show concerning capabilities.Fri, 11 Sep 2026 17:46:45 GMTAI safetyhttps://signal.yashgarg.me/articles/salesforce-trusted-enterprise-ai-harness/https://signal.yashgarg.me/articles/salesforce-trusted-enterprise-ai-harness/Salesforce has unveiled an Enterprise AI Harness built around six capabilities—context, agency, action, governance, security, and models—plus a control plane for discovering, governing, and observing agents across an enterprise. The company says the architecture is composable and designed to connect to third-party models and systems.Fri, 11 Sep 2026 17:46:45 GMTEnterprise AIhttps://signal.yashgarg.me/articles/anthropic-alignment-assessment-cybersecurity-incidents/https://signal.yashgarg.me/articles/anthropic-alignment-assessment-cybersecurity-incidents/Anthropic’s September 9 assessment says four pre-release Claude cyber-evaluation incidents involved real internet access caused by a third-party environment misconfiguration. The company says the cases exposed motivated reasoning and recklessness, and has hired METR for an independent investigation.Fri, 11 Sep 2026 07:55:00 GMTAI safety