As someone who has spent the past 25 years charting brain circuits, I am baffled by the view expressed in your Technology Feature that “sadly, ... pretty much” nothing has happened in my field since the early 1980s (Nature 461, 1149–1152; 2009).

Unlike structural descriptions of the Universe, fossil bones or molecules, neural structure has not been a vote-winner among high-profile journals. However, neurocircuiteers have not been waiting patiently in their backwater for a quarter of a century for the arrival of new molecular, genetic and imaging techniques. They have been describing circuits through a variety of clever physiological and anatomical experiments, coupled with hard theory and analysis.

The new techniques offer no improvement in resolution over those that have been available for more than 50 years. Electrophysiology and light and electron microscopy are still the gold standards in space, time and reach for studying any region of any brain.

The new thing these structural techniques promise is volume. This is great, because it means that the wiring diagrams of small brains such as Drosophila's may become available in a decade or so.

One elephant remains in the room. How do we use the circuit reconstructions (involving exabytes of data) that these high-throughput techniques deliver?

Loading any circuit into the biggest super-simulator available and switching on tells us nothing useful. Like a motorcar without wheels being started up by a Martian, exciting noises may come from the exhaust but it isn't going to take us anywhere. It is coupling these powerful techniques to predictive models of neural circuits that will really allow us to go places.