Filtered for water in history

12.40, Monday 22 Jan 2018


In 2003, the artist Natalie Jeremijenko built a robotic goose.

The robotic goose interface allows people to approach the birds, follow them closely and interact in a variety of ways that would not otherwise be possible without this interface.

The remote-controlled, human-haunted goose can speak:

The goose drivers can ‘talk to’ the geese, issuing utterances through the robotic interface, delivering prerecorded goose ‘words,’ their own vocal impersonations, or other sounds (such as goose flute hunting calls). Each utterance via the robotic goose triggers the camera in the robot’s head to capture 2-4 seconds of video recording the responses of the actual biological geese. These video samples upload to the public web-based goosespeak database that the participants can annotate, i.e. “the goose was telling me to go away,” “he was saying Hi.” As this database of goose responses accretes, redundancy and correlations in the annotations may provide robust semantic descriptors of the library of video clips.

Speaking to animals not by attempting to translate, but by entering their world.

Here is a video of Jeremijenko’s project.

See also: Umwelt, the semiotic world of the organism, including all the meaningful aspects of the world for any particular organism. A bat will see and inhabit a different world than a human, than a goose, than an artificial intelligence.

See also: Timo Arnall’s film Robot readable world, an experiment in found machine-vision footage, exploring the aesthetics of the robot eye.


From 2007, 3D Mailbox is an email client that visualises your messages as a virtual reality 3D environment.

The BBC covered the application, a virtual Miami Beach, a place where your e-mails are represented as bikini clad avatars.

When you receive an e-mail the avatars appear at the entrance to the hotel’s pool. Your spam manager acts a gatekeeper, only allowing through genuine messages to lounge by the pool.

Spam is dispatched to the beach, where they are represented as overweight tourists. Hitting delete results in the spam taking a deadly swim in shark infested waters.

Meanwhile your unread e-mails perform lengths in the pool. Read e-mails tan themselves on the sun loungers.

There are pictures.


It used to be, when I went hiking, I’d step across where a stream was marked on a map or mentioned in the guide, and it would be a measly trickle or a torrential river. Well that’s clearly not a stream, I’d think.

For me a stream is a permanent feature of the landscape. It’s a constant, regardless of rainfall or time of year. It flows gently and runs in a ditch maybe a metre wide. It is probably clear.

It turns out this is rare.

My conception of a stream is a chalk stream.

Underground chalk acts as an aquifer, and then

Bubbling up out of thousands of springs, chalk streams collect first in ponds and then ripple over gravel beds through chalk hills towards the sea. They are typically wide, shallow and crystal clear, their alkaline waters pure thanks to the constant purifying and filtering in the chalk.

They are an irreplaceable relic of our past, created as the ice sheets retreated from England 10,000 years ago.

The temperature remains static.


There are at least 210 chalk streams in the world, and 160 of those are in England.

In particular, in the south of England. And in particular, in particular: where I grew up.

I never realised that my picture of a stream isn’t how most people will think of a stream. Who knew.


Stafford Beer was the cyberneticist and business management pioneer who, in the early 1970s, built Project Cybersyn for the revolutionary government of Chile. Command economy meets socialist proto-internet.

In the early 1960s, he was running a more esoteric experiment, in pursuit of his desire to build an automated factory.

From historian Andrew Pickering’s essay, The Science of the Unknowable: Stafford Beer’s Cybernetic Informatics:

The T- and V-machines are what we would now call neural nets: the T-machine collects data on the state of the factory and its environment and translates them into meaningful form; the V-machine reverses the operation, issuing com- mands for action in the spaces of buying, production, and selling. Between the T- and V-machines lies the U-machine – the homeostat, or artificial brain – which seeks to find and maintain a balance between the inner and outer conditions of the firm

The U-machine.

By the way,

The cybernetic factory was not pure theory. By 1960 Beer had at least simulated a cybernetic factory at Templeborough Rolling Mills, a subsidiary of his employer, United Steel

It is a core tenet of (early) cybernetics that sufficiently complex learning systems are somewhat equivalent, whether they are made of flesh and blood, or vacuum tubes. It is this tenet which allowed the audicity of the cyberneticists to consider building “intelligent” machines, or to model the brain as a network of moving information.

And sure enough, when I went to the library to consult Beer’s collected papers, How Many Grapes Went into the Wine: Stafford Beer on the Art and Science of Holistic Management, Beer discusses the search for his ideal U-machine:

a self-organizing system need not have its circuitry designed in detail – otherwise what virtue is there in the self-organizing capability? Furthermore, if systems of this kind are to be used for amplifying intelligence, a fixed circuitry is a liability. Instead we seek a fabric that is inherently self-organizing, on which to superimpose (as a signal on a carrier wave) the particular cybernetic functions that we seek to model

And he continues:

Dr Gilbert, who had been trying to improve the Euglena cultures, suggested a potent thought. Why not use an entire ecological system, such as a pond?

So Stafford Beer captures a woodland pond, and attempts to train it to run a factory:

Accordingly, over the past year, I have been conducting experiments with a large tank or pond. The contents of the tank were randomly sampled from ponds in Derbyshire and Surrey. Currently there are a few of the usual creatures visible to the naked eye (Hydra, Cyclops, Daphnia, and a leech); microscopically there is the expected multitude of micro-organisms. In this tank are suspended four lights, the intensities of which can be varied to fine limits. At other points are suspended photocells with amplifying circuits which give them high sensitivity.

The intention was to communicate information about the factory into the pond via optical couplings. Earlier attempts, reported by Pickering, included attempts to induce small organisms – Daphnia collected from a local pond –to ingest iron filings so that input and output couplings to them could be achieved via magnetic fields.

The state of this research at the moment is that I tinker with this tank from time to time in the middle of the night.

I have this picture of Beer, in his slippers in his basement, trying to figure out not only how to speak to this tank of water and algae in its own language, but attempting to put it through business school.

What would be the management style of such a factory foreman? Risk averse? A deep sympathy with the principles of sustainability and the circular economy? (Given it sits in a closed-system tank.)

Our modern efforts into machine learning and artificial intelligence have a familiar feel: we place the neural network at the heart of the system… and just turn it on. And although we can’t tell how the neural network recognises a face or optimises a system, we can tell that they have some natural politics: AIs are unable - or unwilling - to correct for their implicit racism and sexism.

What is the umwelt of a pond? What is the umwelt of an AI?

Uber’s marketplace and Facebook’s newsfeed are run by captured artificial intelligences – unreasonably efficient optimisers, blind to human feelings, natural free market libertarians; a warp core of tremendous ability and held only just in check. We don’t know how these things make their decisions, but we are beginning to see the biases in their actions.

Obviously Stafford Beer’s experiments came to nothing: the factories of China are not run by captured, semi-sentient woodland ponds.

Or. Who knows. Maybe we should put one in charge of Facebook.

More posts tagged:
Follow-up posts: