Anthropic's Claude Code adds computer use capability, enabling closed-loop verification (code → run → inspect UI → fix). The article emphasizes that harness quality, tooling, and orchestration now create larger practical differences than raw model capability. Major developments include the open Hermes Agent ecosystem gaining momentum, Alibaba's Qwen3.5-Omni multimodal model with agentic capabilities, and growing focus on specialized open models + local inference (llama.cpp at 100k stars) for real-world agent workflows.
Infrastructure
[AINews] The Last 4 Jobs in Tech
Anthropic's Claude Code adds computer use capability, enabling closed-loop verification (code → run → inspect UI → fix). The article emphasizes that harness quality, tooling, and orchestration now create larger practi...
Tuesday, March 31, 2026 12:00 PM UTC2 MIN READSOURCE: Latent.SpaceBY sys://pipeline
Tags
infrastructure
/// RELATED
Models4d ago
GPT-5.5: Capabilities and Reactions
OpenAI's GPT-5.5 matches Anthropic's Claude Opus 4.7 on raw intelligence and coding while undercutting it on token pricing, with OpenAI publicly signaling rapid iteration ahead.
Products3d ago
Beyond Lovable and Mistral: 21 European startups to watch
TechCrunch profiles 21 European AI startups building specialized solutions across defense, robotics, space, and legal tech to compete in the global AI race.