Please confirm you are human

This browser or connection looks automated. Press and continuously hold the control for 3 seconds to enable Google-hosted web results and, when separately allowed, AI-assisted answers.

A successful check enables 100 search requests. Interactive access does not authorize scraping, systematic collection, or reuse of search output.

Hold with a pointer, or hold Space or Enter.

Web

News

Please enter a web search for web results.

News

Web
The New Stack
thenewstack.io > ai-agents-retrieval-engineering

AI agents are making retrieval engineering a core engineering discipline

1+ hour, 45+ min ago   (17+ words) AI agents are making retrieval engineering a core discipline. Discover how better context drives smarter autonomous decisions....

The New Stack
thenewstack.io > building-ai-agent-harness

Your AI agent is only as good as the harness around it

2+ hour, 45+ min ago   (959+ words) Build production-ready AI agents with effective guardrails, tool contracts, permissions, and tracing systems beyond simple demos....

The New Stack
thenewstack.io > ibm-granite-reasoning-models

IBM's new Granite 4.2 models add reasoning and stay dense

4+ day, 22+ hour ago   (50+ words) IBM’s Granite 4.2 sticks with decoder-only models, adds a 512,000-token context window, and trains its larger versions for agentic work....

The New Stack
thenewstack.io > real-time-ai-scale

Why real-time AI at scale is so hard

1+ week, 1+ hour ago   (623+ words) Discover why real-time AI fails at scale: the failure patterns that surface as your app ramps up, and how to avoid them....

The New Stack
thenewstack.io > upstage-solar-pro-4

"Save frontier models for frontier problems": Why Korea's Solar Pro 4 is a workhorse agent reliability play

1+ week, 3+ day ago   (369+ words) Upstage AI's Solar Pro 4 targets reliable AI agents for enterprise workflows, promising 90% lower costs than frontier models for document tasks....

The New Stack
thenewstack.io > glm-5-3-anthropic-distillation

An industrial-scale distillation of models, or subtle benchmaxxing: What developers really think of GLM-5.3

1+ week, 4+ day ago   (536+ words) Z.ai's GLM-5.3 claims big coding gains, but AI professionals say the real story may be distillation from Anthropic's Claude models, or a case of subtle bencmaxxing, not novel training....

The New Stack
thenewstack.io > amodei-open-weights-compute-regulation

"Open weights are nowhere near a sufficient solution": Dario Amodei fires back on AI power

1+ week, 6+ day ago   (511+ words) Dario Amodei says open weights just shift AI's power concentration to whoever controls the most compute, and argues for tiered safety rules instead....

The New Stack
thenewstack.io > glm-5-3-post-training-coding

GLM-5.3 didn’t change the base model — where did its coding gains come from?

2+ week, 2+ day ago   (30+ words) Z.ai's GLM-5.3 reuses the GLM-5.2 base model and gets all its gains from post-training. Benchmark jumps are large, but the weights won't drop for two weeks....

The New Stack
thenewstack.io > ai-pipeline-token-optimization

Why your AI pipeline costs 10x more after the demo

2+ week, 3+ day ago   (206+ words) Stop overpaying for AI in production. Master key architectural token optimization strategies to lower costs and boost speed....

The New Stack
thenewstack.io > manus-meta-data-deletion

Your AI agent remembers everything. Here's what happens when its owner changes.

2+ week, 4+ day ago   (27+ words) Manus will delete some user data as it separates from Meta after China blocked the $2B deal. Here's what developers need to save — and when....