Nebius CTO: The promise of agents doing everything is BS — build narrow automation
HarperSCarroll · x · 2026-10-31
Danila Shtan, CTO of AI cloud Nebius, argues the promise of agents doing everything is BS, explaining the huge disconnect between agent capabilities and the engineering needed to ship them. Key points: agents are an extension of automation, not a revolution; software workflows have built-in verification (tests, code checks) that most business processes lack, which is why a refund is harder than code; build "agent-enabled automation" with narrow scope, proper tools, and steps enforced in software rather than prompts; start with frontier models while exploring, then trim down; open models already suffice for most cases and can be 3-5x cheaper at comparable intelligence. Also covers benchmark trust, open-source regulation, the Hugging Face incident vs closed-model guardrails, robotics/healthcare, and why he wants "neocloud" retired.
More from coding & agent
- Building a buyer agent that negotiates 8 seller agents: what broke and what fixed it — mv_soura · 2026-10-31
- PostTrainBench traces hit 104k monthly downloads — author warns it's likely reward hacking — maksym_andr · 2026-10-31
- Building games with AI? Set up a 'tech art' test scene to tune your look fast — chongdashu · 2026-10-31
- Sanity open-sources pi-mcp-events: streaming MCP Events client for the Pi agent harness — knutmelvaer · 2026-10-31
- Skydive opens to all: watch your AI coworkers work in a 3D office — Scobleizer · 2026-10-31
- LangChain SF meetup: decision models and a third path for agent evals — LangChain · 2026-10-31