Inside Ramp's Quiet, Profitable AI Rollout
The interesting story is not the demo. It is the second month, when RAG-as-a-service either earns its keep or gets quietly rolled back.
Yuki Tanabe3 min read
The archive
All articles
100 stories from the Inference Daily desk.
Why analysts Are Suddenly Standardizing on Notion AI
Beyond the launch posts, Llama 4 is reshaping how small studios approach customer support. We talked to the people actually using it in production.
Yuki Tanabe3 min read
Shipping AI You Can Defend: A QA Field Report
The interesting story is not the demo. It is the second month, when fine-tuned distillation either earns its keep or gets quietly rolled back.
Mira Castellanos3 min read
Inside Spotify's Quiet, Profitable AI Rollout
Duolingo is the latest in a string of teams treating tool-first agents as the default, not the experiment. Here is what they got right — and what they are still figuring out.
Mira Castellanos3 min read
When to Replace research synthesis With an Agent — and When Not To
Intercom is the latest in a string of teams treating long-context workflows as the default, not the experiment. Here is what they got right — and what they are still figuring out.
Mira Castellanos3 min read
Inside a Production Agent That Actually Ships With small-model orchestration
Beyond the launch posts, DeepSeek V4 is reshaping how researchers approach incident response. We talked to the people actually using it in production.
Yuki Tanabe3 min read
The Post-Link Web: What agent mode Means for Publishers
Duolingo is the latest in a string of teams treating evals-first development as the default, not the experiment. Here is what they got right — and what they are still figuring out.
Priya Raman3 min read
Anthropic Announces Claude 4.5 Sonnet, and the Frontier Moves Again
The interesting story is not the demo. It is the second month, when RAG-as-a-service either earns its keep or gets quietly rolled back.
Jonas Halvorsen3 min read
LLMs·
Cohere Trims the Stack: Mistral Large 3 and the Cost Curve
Duolingo is the latest in a string of teams treating long-context workflows as the default, not the experiment. Here is what they got right — and what they are still figuring out.
Mira Castellanos3 min read