Agentic AI
AI that pursues a goal across multiple steps — choosing tools, acting on the world, and adjusting based on results — rather than answering a single prompt and stopping.
Learn first
A standard language model takes a prompt and returns text. An agentic system keeps going: it breaks a goal into steps, calls tools (web search, code execution, file writes, APIs), reads the results, and decides what to do next. The loop runs until the task is done or the agent gets stuck. The hard part is reliability over many steps — each decision is a new chance to drift, misread context, or compound an earlier error. Most production failures are not model failures; they are failures of the loop.
Where it came from
In megatrends
Related players
Finds citing this concept
Overall = breadth + depth + substance + ½·novelty + freshness (a small boost that halves each week)
Claude Mythos
In one month, fifty partners pointed Anthropic's bug-hunting model at production code and reported more than 10,000 serious vulnerabilities.
Project Glasswing
Anthropic's most capable model found thousands of unpatched zero-day flaws in its first weeks — so the lab is routing it to defenders before shipping it to everyone.
ChatGPT, your bank teller
ChatGPT Pro users can now wire up real accounts across 12,000 financial institutions through Plaid and ask, in plain English, whether they're on track to buy a house.
Design Conductor
Handed a 219-word spec, an AI agent designed a working RISC-V processor and carried it alone all the way to a tape-out-ready layout in twelve hours — no engineer in the loop.
gstack
Y Combinator's CEO published the exact Claude Code setup he codes with — slash commands that each play an engineering role — and it out-starred most of GitHub's real software.
How this connects
Tap a node to open it