Amazon's Connect Talent: AI Agents Automate Mass Job Interviews

Amazon has released Connect Talent, a new AI hiring software designed to automate mass job interviews for large-scale recruitment, such as holiday seasonal hiring. The tool is part of Amazon's broader investment in autonomous AI agents — software that can plan, decide, and act with little to no human intervention.
Key details
- Connect Talent can conduct AI-led interviews 24/7, screen candidates, and prepare notes for recruiters. No human involvement in the interview loop.
- Amazon hired around 250,000 seasonal workers last year; the software is built for similar scale.
- Candidates will be informed they are being screened by AI. Colleen Aubrey, AWS SVP of applied AI, acknowledged the voice interaction is still being refined to sound convincingly human.
- Amazon also introduced Connect Decisions, an AI agent that analyzes and compiles supply chain and purchasing data to equip planners with actionable data.
- The company's new AI design philosophy called "humorphism" aims to humanize AI — adapting AI behavior to how humans work, rather than forcing humans to adapt to AI.
Context on AI agents
The announcement comes at a time when Alphabet, OpenAI, and Anthropic are all pushing deeper into enterprise AI agents. Amazon has cut roughly 30,000 corporate jobs since October, with some tied to AI-driven efficiencies.
The Connect Talent roll-out is a concrete example of AI agents being used for high-volume operational tasks — but also raises questions about bias, candidate experience, and oversight in automated hiring pipelines.
📖 Read the full source: HN AI Agents
👀 See Also

Claude Code v2.1.203: Major Bugfix Release with Background Session Recovery and Performance Improvements
Claude Code v2.1.203 fixes background session stalls on macOS, token expiration recovery, PATH inheritance issues, and reduces binary size by 7 MB with startup memory savings.

Diagnosing Operational Drift and Task Amnesia in OpenClaw with Gemini 2.5 Flash on Proxmox
OpenClaw users report issues with persistent workflows on a Proxmox VM, citing operational drift and task amnesia. Despite stable performance in one-off tasks, the Gemini 2.5 Flash model struggles with automation and memory in this setup.
Discovered Materials: AI Agents Discover 500+ New Materials, But Only 1 Has a Plausible Synthesis Route
Discovered Materials (YC P26) used frontier LLMs to computationally discover 500+ new semiconductor materials. Only 1 has a plausible synthesis route, highlighting the lab-to-fab gap.
Opus 4.7 Can Follow ~500 Instructions, Up from ~150 a Year Ago
Research updated in May 2026 shows Opus 4.7 can reliably follow ~500 instructions, compared to ~150 in July 2025. GPT-5.5 handles ~5000. Implications for CLAUDE.md file size.