[AINews] The Last 4 Jobs in Tech

2026-03-31 · Source: Latent.Space - Www.latent.space · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Advanced, quick

Summary

This intelligence brief, covering March 28-30, 2026, highlights significant advancements in AI, particularly in agentic workflows, multimodal models, and local inference. Anthropic introduced computer use within Claude Code for Pro/Max users, enabling closed-loop verification for code development. OpenAI released a Codex plugin for Claude Code, signaling a shift towards composable coding stacks. The open-source Hermes Agent received a major update, featuring multi-agent profiles and an expanding ecosystem for traces and remote control. Alibaba launched Qwen3.5-Omni, a multimodal model with native text, image, audio, and video understanding, supporting 113 speech-recognition languages and 36 spoken languages. Local AI also saw a milestone with llama.cpp reaching 100,000 GitHub stars, emphasizing the growing importance of portable, non-vendor-locked infrastructure and specialized open models.

Key takeaway

For CTOs and VPs of Engineering evaluating AI infrastructure, the shift towards composable agent harnesses and specialized open models demands a re-evaluation of your current strategy. Prioritize investing in flexible, non-vendor-locked local inference solutions and explore multi-agent architectures to optimize development workflows and reduce operational costs, as demonstrated by Shopify's $5.5M to $73K/year reduction.

Key insights

AI development is rapidly advancing towards composable agentic systems, multimodal capabilities, and efficient local inference.

Principles

• Coding stacks are becoming composable harnesses.
• Harness quality is a first-order variable for agent performance.
• Specialized open models are a key deployment pattern.

Method

The CAID paper proposes centralized asynchronous isolated delegation for SWE agents, using manager agents, dependency graphs, isolated git worktrees, self-verification, and merges to improve performance.

In practice

• Utilize Claude Code's computer use for closed-loop code verification.
• Explore Hermes Agent for multi-agent profiles and open ecosystem tools.
• Consider specialized open models for proprietary data.

Topics

• AI Agent Architectures
• Multimodal AI Models
• Local AI Inference
• AI Harness Engineering
• Open-Source AI

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Engineer, Machine Learning Engineer, AI Scientist

Related on AIssential

• not much happened today — Restricted cyber-capable AI models, open agent architectures, and efficient local inference are shaping the AI landscape.
• How Harness-as-a-Service Will Change Agents — Harness as a Service is democratizing AI agent development by providing pre-built runtime environments.
• not much happened today — AI development is rapidly converging on composable agentic systems, multimodal understanding, and efficient local inference.
• 🔮 Jensen’s OpenClaw thesis — The AI economy is rapidly shifting from training-centric to inference-centric, demanding specialized hardware and organizational strategies.
• [AINews] The Inference Inflection — The AI industry faces an "inference inflection" driving massive compute demand and platform evolution for coding agents.
• Harness Engineering for Business Operations: What Cinema Teaches About Agent-First Systems — Inverting traditional approaches by imposing extreme constraints can drive significant engineering breakthroughs.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Latent.Space - Www.latent.space.