Sign up to our newsletter

Welcome to See Through News

Speeding Up Carbon Drawdown by Helping the Inactive Become Active

[wpedon id=3642]

AI’s Dirty Secret, or How To Spend Half A Million Dollars of Supercomputing – Omnibus Edition.

AI dirty secret carbon drawdown sustainability global south cloud computing competition nigeria mexico

All 10 Episodes of Series 1 of The See Through News Podcast, which is entirely true.

This omnibus edition features all 10 Episodes of this entirely true suspense thriller concerning cloud computing, the Global South and working miracles on zero budget.

Like all stories, it’s best to start at the beginning:

Episode 1 – Monkeys Climbing Trees

Episode 2 – AI’s Dirty Secret

Episode 3 – A Surprise Donation

Episode 4 – Best Christmas Wrapping Present Ever

Episode 5 – An Unusual Introduction

Episode 6 – Rules, Referees and Global South Haystacks

Episode 7 – Golden Needles

Episode 8 – Tick Tock

Episode 9 – Did It Work?

Episode 10 – What Happened Next?

(to receive notifications as soon as new episodes are released, subscribe to the See Through News YouTube channel or your preferred podcast platform).

Written, produced & mixed by Robert Stern

Narration & music by George Hinchliffe

If you enjoyed this series, why not try these fictionalised true stories from our sibling podcast The Truth Lies in Bedtime Stories, from See Through News.

The Truth Lies in Bedtime Stories is a See Through News production.

See Through News is a non-profit social media network with the Goal of Speeding Up Carbon Drawdown by Helping the Inactive Become Active.

For more visit seethroughnews.org

Transcript

Episode 1: Monkeys Climbing Trees To Reach The Moon

Imagine a student flat, in Scotland, in the mid-1980s.

Now picture an odd couple of flatmates, studying odd things. 

Canadian Mark Drummond, is doing a PhD in Artificial Intelligence. This is partly because his parents are Scottish, but mainly because Edinburgh has the only AI Department outside North America. Mark is fascinated by using technology to change how we live.

Anglo-Australian Robert Stern started studying Politics & Philosophy. When he discovered Edinburgh was the only Scottish university with a Chinese Department, he switched courses. Robert is fascinated by how humans live with, or change, the environment.

Robert reckons AI is voodoo magic. He loves hearing Mark try to explain it. 

Mark finds Chinese voodoo magic. He loves hearing Robert try to explain it.

Over the kitchen table, Mark outlines an ongoing debate between AI pioneers. Some believe robots can, in principle, one day emulate human intelligence. Others reckon it’s like monkeys climbing trees to reach the moon – they may make progress, but not meaningfully.

In response, Robert grabs a pen and paper to tell Mark a classical Chinese dirty joke he’s just stumbled across. It’s just as vulgar as you’d expect a joke about a lusty monk and a winsome nun to be, but it requires writing Chinese, and dates back to the Tang dynasty.

**

Mark and Robert graduate. Their paths diverge, but they stay in touch. 

Gradually, events make their obscure study choices look increasingly smart. 

As computing power grows, AI monkeys ascend further up trees. Mark’s AI doctorate becomes a hot property. 

As China grows, the ‘sleeping giant’ awakes. Robert’s Chinese skills become more employable.

Mark works on the Mars Rover at NASA. He joins the team that made SIRI and sells it to Apple. He becomes a serial dotcom entrepreneur at the cutting edge of real-world AI applications. He thrives on both sides of the Silicon Valley venture capital pitching table.

Robert trades textiles at a Japanese multinational, then quits to move to Tokyo. Now fluent in formal office Japanese, he touts himself as an environmental consultant to major Japanese corporations. He fails, falls into TV news, produces and reports for American network news in Tokyo, Hong Kong, Shanghai and Beijing. He moves home to Britain to make documentaries, teach filmmaking and journalism, broadcast, and consult for international co-productions, sports events and TV entertainment formats.

Throughout, Robert & Mark stay in touch, still fascinated by each other’s field of study. 

Throughout, China and AI continue their ascent from obscurity to dominance.

Throughout, Robert’s 80s concerns about the human impact on the environment grows from an eccentric hobby to become a focus on an existential crisis.

**

By 2021, Mark heads an AI project at Apple so secret his immediate boss isn’t allowed to know what it is. If we told you what it was, we’d both either probably be dead, or very senior Apple executives.

In 2021, Robert abandons paid work to set up a social media network called See Through News, applying storytelling tricks to carbon reduction. Thirty years after failing to make up a paid job as an environmental consultant, Robert succeeds in making up an unpaid one. The See Through News Goal is Speeding Up Carbon Drawdown by Helping The Inactive Become Active. 

He and a group of volunteers create a series of eye- and ear-catching projects: a newsletter, a podcast, community filmmaking, Facebook Notice Boards, even an AI project.

They appear unconnected, but are actually different starting points to the same destination of measurably reducing carbon. Each project is designed to appeal to some group or other of ordinary people as a point of engagement.. 

It was called See Through News for two reasons. 

One, its unconventional methodology, while convoluted, was transparent. Like a conjurer explaining his tricks as a DVD Extra, every manipulation and trick was explained, to the curious. 

Two, see-through things don’t signal their purpose. Having observed that many climate activist organisations with words like ‘Green’, ‘Sustainable’ or ‘Carbon’ in their titles deterred some ordinary people as soon as they saw these names, Robert thought it worth trying to first engage people with activities that were appealing in their own right, before revealing any carbon-reducing purpose. 

Most people recognise the science and reality of climate change, but feel powerless to do anything about it. See Through News deployed ‘green’ trigger words as little and as late as possible, ideally not at all.

As well as the school projects, film screenings and video games, Robert regularly writes journalistic articles offering climate perspectives on hot topics, which he publishes every Sunday in the See Through News newsletter.

**

In 2022, Large Language Models finally thrust AI from backstage obscurity. Blinking in the limelight, ChatGPT emerges as AI’s breakout star.

Thanks to Mark and other AI student university friends, Robert has tracked AI developments since the 1980s, but is frustrated by the gap between what his AI chums tell him, and what appears in news headlines. 

As a TV reporter, Robert is used to condensing complex stories into two-minute reports, so knows all too well why mainstream media thrusts AI into the binary boxes of Good Robot Story and Bad Robot Story. 

Now as a climate activist, Robert reckons mainstream media isn’t providing the deeper nuanced understanding of AI’s actual technology that we should require in order to arrive at an informed opinion on this hot topic.

To bridge the gap, Robert writes an article for the newsletter. Recalling the phrase Mark used in that student flat to describe AI in its 1980s infancy, he calls it AI, Climate Change & Monkeys Climbing Trees To Reach The Moon.

In his research, Robert comes across a cloud computing management startup. UK-based YellowDog claims to offer the greenest form of AI processing available. Robert smells greenwash. 

He contacts YellowDog to challenge their claim. Tom Beese has just taken over as CEO of YellowDog, in addition to heading four other bleeding-edge hi-tech startups, but not only takes Robert’s call, but welcomes his interest. 

Over many long phone calls, Tom patiently explains exactly how YellowDog’s advanced mathematics and cutting-edge tech minimise computing’s carbon footprint. Tom’s contributions, and imaginative analogies involving football stadia, form a key part of the article. 

It’s not all business, Tom is curious about See Through News and Robert’s journey to founding it. Robert is intrigued by Tom’s previous lives as a lawyer, army captain, novelist and boss of a virtual semiconductor fabricator. 

The business bit is shocking, just not in the way Robert anticipated. He thought he knew a thing or two about both AI and computing, but is appalled by what Tom now tells him.

In Episode 2, AI’s Dirty Secret, Robert keeps digging, and discovers a dirty secret.

Episode 2: AI’s Dirty Secret 

Robert started writing Monkeys Climbing Trees To Reach The Moon to share with a general audience his amateur understanding of how AI actually works.

He ends up being shocked at his own ignorance, as a climate activist, at the scale of computing’s carbon footprint.

The more detail Tom goes into about YellowDog’s clever carbon-reducing tech, the more its significance sinks in. That innocent-sounding, ethereal, floaty concept of ‘cloud computing’, it turns out, actually takes place in very much earthbound data centres. 

These data centres are usually hidden from public view, but they power the Internet. They also consume huge amounts of energy just to run idle. 

An average data centre requires enough power to run a small town, just to keep their processors from bursting into flames. To avoid turning the lights out in nearby towns, many have their own power stations.

As a journalist, and now an activist focused on carbon emission reduction, Robert is astonished at both the scale of computing’s dirty secret, and by its absence from any of the breathless discussions of AI now dominating the mainstream media discourse. 

The Good Robot articles skip over this inconvenient truth. If AI is going to ‘save the world’, why bring up the awkward and distasteful matter of its insatiable appetite for energy?

Even the Bad Robot articles focus on futuristic Hollywood visions of robots taking over, enslaving or eliminating humans. The only article Robert can find sounding the alarm about the exponential rise in the Internet’s carbon footprint is in a specialist publication for the engineers who design and manufacture the chips that go inside the data centres.

Just as extreme weather events have climate scientists who’ve been predicting them for years banging their heads in frustration, chip engineers can’t believe everyone is still ignoring reality. 

Like climate change itself, the issue of the Internet’s growing carbon footprint is no longer a theoretical future problem, but a very real current threat. When chip designers meet, they show each other graphs showing their own hockey stick curve. 

Instead of concentrations of greenhouse gases in the atmosphere, this sudden sharp jump shows the increase in computing power demand created by ChatGPT and all its imitators.

AI is turning out to be the computing equivalent of the invention of the internal combustion engine – the impetus for a sudden, massive increase in carbon consumption, just when we need to start slashing it to zero and beyond.

And what critical new demand is driving this eating up of our already in-the-red carbon allowance? Convincing, but imaginary pictures of celebrities doing unlikely things, deep fake porn and disinformation, and exam cheating, that’s what.

‘Is this what we’re going to tell our children we blew our carbon allowance on?’, chip engineers ask each other when they meet. ‘Data centres sucked up all our available energy, cancelling out any carbon reductions we made elsewhere, for pictures of the Pope in a funny hat? ‘

It was easier for Robert to understand why his AI friends ignore this unpleasant reality, easier still why venture capitalists eager to cash in on the Next Big Thing, decline to bring it up. After decades climbing the tree, the monkeys wanted to have their day, even if they’ve not yet reached the moon.

**

Like most climate activists, and increasingly the public, Robert is hyper-aware of the carbon profligacy of, for example, the aviation industry. The press is all over it – flight-shaming is clickbait catnip. All those column inches and air time, branding celebrities, politicians and activists as hypocrites for flying to a climate protest or environmental conference.

Given all the hot air, it’s almost disappointing to learn that aviation accounts for… 2% of total global carbon emissions. 

If this number appears at all, it’s usually buried very low in the finger-pointing articles, and you can see why. Two percent of anything seems like a trivial thing to spend so much time getting so worked up about.

For activists, this is a tricky issue. Treating long-haul flights like local buses is a great symbol of our carbon profligacy, and need to adapt to a sustainable lifestyle. But for results-focused activists like Robert, stoking up a highly divisive Twitterstorm about something that only generates 2% of emissions also looks like a very effective tactic for vested interests who are merrily pumping out the other 98%.

In this light, flight-shaming starts to look like a classic ‘dead-cat’ PR trick – if you want to prevent something awkward being discussed, throw a dead cat on the table. It usually changes the subject.

Meanwhile, as Tom is patiently teaching Robert, all our apparently harmless, paper-free online activities – all those video calls, streaming services and infinite and eternal access to cat videos, are already generating more than twice as much carbon as flying. 

Worse still, Tom explains, the vast majority of the chips used in the data centres that store and process all this online activity are still designed to optimise performance, not reduce power consumption. 

Computing power consumption is already at least 4.5% of global emissions, but the Large Language Model breakthrough being celebrated in all those articles about ChatGPT and its imitators, will supercharge demand even further. 

In a series of further, long phone calls, Tom explains the tech, science and data behind this dirty secret. 

Once he’s got his head round it, Robert publishes Computing’s Carbon Footprint – the Other AI Threat.

In Episode 3, A Surprise Donation, Robert receives an unexpected call from Tom, just before Christmas. 

Episode 3: A Surprise Donation

2022 has been a busy year. By December 15th, Robert is ready to wind down for the Christmas holidays. 

He’s thinking it may be time to take a breath, and reflect on the project that’s taken over his life.

There’s much on which to reflect. 

It’s nearly two years since Robert abandoned earning a living as a documentary filmmaker, broadcaster and consultant to set up See Through News, a zero-budget climate activist network. 

This decision gave Robert a purpose, clarity and meaning that is shared by a growing international network of volunteers. It may also have hastened the end of his 25-year-old marriage.

As the See Through News network grows, its appeal is becoming more clearly defined. The characteristics that set it apart from most environmental activist organisations are getting easier to explain, and demonstrate, as more projects are developed and added to the increasingly diverse roster.

One such project is developing so strongly he thinks it might need its own website, brand and identity. See Through News’s focus on measurably reducing carbon has thrown light on a rather important problem – measuring carbon.

‘Carbon auditing standards’ might not sound very sexy, but measuring things properly is an essential starting point to claiming any progress. 

The deeper Robert delves into the current world of carbon auditing standards, the more convinced he becomes that the chaotic, expensive, competing commercial standards are unfixably unfit for purpose, benefiting greenwashers and hence penalising companies who genuinely seek change.

Robert is talking to some carbon auditing expert volunteers about developing a radically new standard that’s accurate, free to use, transparent and open source. It’s early days, but they’re getting enthusiastic feedback so far, especially for emulating the See Through News zero-budget model. As a journalist, Robert’s first instinct is to Follow The Money, so he knows the power of not having any money to follow. 

Robert is thinking of calling this spin-off See Through Carbon, with the tag-line, ‘If you can’t buy integrity, why should you be able to sell it?’.

**

After twenty years running a production company in what he likes to refer to as the ‘beggingest business in the world – independent documentary filmmaking’, Robert is delighted to discover that, so far at least, the advantages of not having a bank account far outweigh any drawbacks.

Operating through volunteers alone not only provides instant integrity, it’s also super efficient. Remove all the time and energy that money-based charities spend on raising, approving and spending money, and you’re instantly ten times more productive, and twenty times more nimble. 

Money-free activism permits a laser-like focus on the mission – for See Through News, this means measurably reducing carbon.

The See Through News combination of crystal-clear goal and refreshingly-original method is attracting activists who share Robert’s frustration whenever he hears the phrase ‘saving the planet’. Like him, they prefer the ambition of ‘mitigating the worst effects of human-induced climate change on human civilisation’. Less catchy, more meaningful.

‘Raising awareness’, isn’t a bad thing, but if ‘educating’ and ‘informing’ is an end in itself, rather than a means, no carbon is reduced as a result. 

**

A year ago, See Through News road-tested some of its ‘Transparent Trojan Horse’ activities at COP 26 in Scotland. 

The team had seen how the world’s most active climate activists  – and ordinary Glaswegians, responded to a street theatre quiz game, a superhero drawing competition and a guerilla gig for ukulele and bagpipes, instead of pamphlets and protests. 

For nearly a hundred consecutive Sundays, Robert had produced a newsletter summarising progress. He was teeing up the Sunday December 18th edition, but the following one would be Christmas Day. 

Nothing too noteworthy had occurred, and even the keenest See Through News newsletter subscribers probably wouldn’t be thinking about effective climate action on Christmas Day. 

Sitting at his keyboard, struggling to come up with an article for Sunday’s newsletter, Robert wonders if anyone would really notice, or care, if he misses one newsletter?

That’s when Tom calls. 

**

Robert wishes him seasonal greetings, they swap holiday plans. As Robert is thinking of winding up, Tom says this isn’t just a social call. As CEO of cloud computing management platform YellowDog, Tom wants to make a corporate donation to See Through News.

Robert gently reminds Tom See Through News has no bank account, and is entirely run by volunteers.

Equally gently, Tom tells Robert he’s aware of that, which is why the donation is in the form of a voucher. More precisely, a ‘free licence’.

Two conditions, Tom says:

One, the voucher has to be used to further the See Through News Goal of ‘Speeding Up Carbon Drawdown by Helping the inactive become active’.

No problem there, says Robert. Hardly counts as a condition.

Two, Tom goes on, there’s a strict deadline. The free licence has been accounted for in YellowDog’s annual budget, and expires at the end of the financial year, on April 10th, less than 5 months away.

Also not a problem, says Robert, breezily. For a former TV news producer, 5 months is an eternity.

So what’s the voucher for, and how much is it worth?

Supercomputer-grade cloud computing, replies Tom.

And half a million dollars.

Well, that solves the problem of what to write about in the pre-Christmas newsletter. Robert writes an article called Our First Donation – Half A Million Dollars Worth of Cloud Computing.

But with offices empty during the Christmas holidays, time ticking down to the suddenly imminent-sounding April 10th deadline, no profile, no track record and no budget, what on earth is See Through News supposed to do with this bolt-from-the-blue gift?

In Episode 4, Best Christmas Present Wrapping Ever, Robert spends his Christmas holidays working out what Tom’s gift is, and how to wrap it.

Episode 4: Best Christmas Present Wrapping Ever

When Tom calls to make his donation, office Christmas party season is in full swing. 

By the time Robert gets his head around the enormity of the donation, most people are, in reality, or in their minds, on holiday.

For the next fortnight, Robert’s only sources will be Tom, online research, and whatever specialist advice and opinion he can glean from friends, acquaintances and See Through News network volunteers between the parties and the family visits. 

First, Robert needs to get a handle on what exactly this gift is. He asks Tom if one or two of See Through New’s projects might be able to make use of this supercomputer-level processing?

Tom tells him straight away no See Through News projects are currently anywhere near requiring the kind of processing power on offer. Cloud processing is not like replacing your old laptop with a new one. It’s the difference between crawling with your hands and feet tied together, and sprinting.

So See Through News’s job is to find projects that serve its ten-word Goal, that can make use of cloud computing.

The first part of its Goal is straightforward. ‘Speeding Up Carbon Drawdown’. ‘Carbon drawdown’ is shorthand for either removing existing carbon from the atmosphere, or reducing the amount currently being transferred there from the ground by burning non-renewable fuels. If Earth is a bath filling with water, carbon drawdown projects help pull out the plug, or turn off the tap. 

In short, Robert and Tom need to find worthy projects that promote sustainability. So far so clear.

The second part of the Goal ‘By Helping The Inactive Become Active’ was less obvious, and the deeper Robert examined it, the murkier the view became. 

Usually, helping the inactive become active meant convincing people who accept the science and reality of climate change, but feel powerless to do anything about it, to take a specific action that could measurably reduce carbon.

By now Robert understands how, even with YellowDog’s mitigations, cloud computing pumps out massive carbon emissions. 

The challenge, then, is to find one or more sustainability-related projects that:

  • have a huge amount of data 
  • need complex processing
  • would be transformed by the injection of supercomputing
  • would result in original research 
  • And would measurably reduce way more carbon than that produced the supercomputing required to generate it

Robert puts these challenges to various AI friends. Some work in academia, others in commerce. They agree on three key points, but these in turn imply three Problems: one Minor Problem, one Moderate Problem, and one Massive Problem. 

**

First, the consensus:

  1. Half a million dollars buys a serious amount of computational horsepower
  2. Commercial businesses would already have budgeted for supercomputing, which leaves academia. Even for free gifts, advanced research projects require the approval of committees that might only meet a couple of times a year. This means the most likely candidates – who might have the volume of data ready, and require the degree of complex processing to make use of this horsepower – are current PhD students. 
  3. April 10th is already a ridiculously tight deadline, and the competition should launch as soon as possible. Realistically, the earliest possible date is January 1st 2023.

So a few days bring Robert one kind of clarity. The best way to give away half a million dollars of supercomputing by April 10th is to use the YellowDog free licence as a prize fund for some kind of Competition targeting PhD students working on sustainability projects.

So far so good. Now for the Three Problems. 

** 

The Minor Problem is what to call it. This was easily resolved. The competition would be called the See Through Carbon Competition.

The principles of the See Through Carbon auditing standard immediately helped shape the ideal candidate’s qualities: 

  • Current carbon auditing standards aren’t working because they’re expensive, inaccurate, opaque and proprietary. 
  • See Through Carbon is free, accurate, transparent and open source.
  • The See Through Carbon Competition would favour projects that reflected those values.

The supercomputing prize would be offered to projects producing research that could rapidly be deployed to measurably reduce real world carbon. Minor Problem solved, with the added bonus of helping establish the Competition’s criteria.

The Medium-size Problem was the April 10th deadline. Everyone working in academia seemed to find it ludicrous. They tell Robert that without any budget, profile, network or track record, See Through Carbon won’t stand a chance. Even those working in the supercharged dotcom business world find the deadline daunting to the point of impossible.

But none of them has worked in TV news production. Robert spent decades meeting impossible deadlines on negligible resources. He’s never organised a zero-budget supercomputing competition before, but he’d been CNN Beijing Bureau Producer for a US Presidential visit to China, and survived in the impecunious world of independent documentary production for two decades. 

It wasn’t obvious to Robert why organising a competition in a few days, during a holiday period, with no money and a crazy deadline is much different from previous obstacles he’s found a way round, over, or under.

The third problem, the Massive one, however, seems so big as to be intractable. 

It’s not that the Competition can’t work, but that it would barely be worth all the effort.

Finding existing sustainability PhD projects in top universities shouldn’t be hard at all. The See Through News network includes academics in leading AI and computer science departments from Edinburgh to London to California, Canada and Japan. 

Off the top of their heads, they can name half a dozen PhD students working on sustainability-related issues, with massive datasets perfectly prepared for cloud processing – but this is precisely the problem.

Awarding projects like these a share of the YellowDog prize fund would at best simply advance them by a few weeks or months. They’ve already budgeted for cloud computing in their proposals. Most are already waiting in line for allocated time slots at their university supercomputers.

Winning the See Through Carbon Competition would merely enable them to queue-jump. This barely seems worth all the hassle.

The solution seems obvious – pass on Tom’s unexpected Christmas gift to researchers who can never dream of being able to access this kind of technology. Find projects for whom supercomputing would be transformative, making possible things that would otherwise be impossible.

This means advertising the Competition to the world’s poorest countries. Or as they prefer to call themselves, the Global South.

But with only a few days before the Competition has to launch to give itself any chance of meeting its deadline, how on earth can such candidates be found? 

And even if the Competition can somehow access such a massive haystack, will it even contain any needles?

In Episode 5, An Unusual Introduction, we learn of the critical role played by a Dutch woman bobbing around in the Atlantic Ocean.

Episode 5: An Unusual Introduction

Robert rises early on Christmas Day 2022. It’s a Sunday, so his mind is already thinking about the newsletter.

In a few hours he’ll be in a Christmas lunch food coma. First, he sits down to write up his dilemma  – how to find sustainability researcher needles in Global South haystacks.

In time for cocktails, he hits Publish, and sends an article titled ‘See Through Carbon Competition – a Half Million Dollar Announcement’ into the ether. Robert can now enjoy the rest of the day without the newsletter hanging over him, but he does wonder if anyone is actually going to read it.

The newsletter, free every Sunday, only has a few hundred subscribers, most of whom would be distracted by the busy Christmas period. Mentioning the dilemma feels like an act of reflective therapy, rather than a realistic appeal for a solution. 

Still, at least it’s some interesting content in a slow news period. 

**

Putting out the newsletter every Sunday has become part of Robert’s routine. Even if its circulation is tiny, and his articles largely ignored, writing original articles on neglected topics every week is good discipline. It keeps him well informed, deepens his understanding of complex problems, and each Sunday further refines the See Through News methodology and mission.

Sometimes the main article is a throwaway climate take on a current topical issue, like the movie Don’t Look Up, dashed off in a couple of hours. Sometimes it’s a deep dive, requiring weeks of research, like one Robert had written a few months before called Dr. Dryden & The Missing Plankton

This was about a parallel existential threat to human civilisation and biodiversity created by burning fossil fuels. Not the blanket effect of carbon emissions on our atmosphere, but how they are gradually increasing the acidity of our oceans.

*

Dr. Dryden & The Missing Plankton was an investigation of the science, and media treatment, of a remarkable citizen-science project. 

In the 1980s Dr. Howard Dryden had turned his marine biology PhD project into a global water treatment business, supplying clean water to everything from entire cities to huge industrial complexes.

Once he started doing business with marine parks, he returned to his original passion, and was shocked by how polluted our oceans had become. 

He sold up his stake in the business, bought an ocean-going yacht, and applied his commercial innovations to the basic science of measuring the seas’ fundamental health indicator – plankton. 

Dr. Dryden invented a new method of taking water samples, and trained a couple of dozen Atlantic yacht sailors in how to use it. They take regular samples, and email the images to a lab Dr. Dryden had set up at an Edinburgh University research hub, where they’d be subjected to expert analysis to count the number of plankton.

They couldn’t find any. 

**

Conventional plankton-counters pooh-poohed Dr Dryden’s findings, either dismissing his technology, or saying he was looking in the wrong places. Mass media, less concerned about such technical nuances, simply mis-reported the results with sensationalist headlines.

Few people read beyond the headlines, including scientists who should have known better. 

A critical scientific question rapidly descended into a Twitterstorm of name-calling. Any scientific reason was drowned out in polarised grandstanding from either side of the social media schoolyard. 

Dr. Dryden & The Missing Plankton was a typical See Through News article. It tried to, well, see through the news, taking the time to explain the science objectively and in detail, and analyse how and why mainstream media didn’t.

You’ve probably not heard anything about this story, which is precisely why Robert spent weeks researching it. After tracking Dr. Dryden down to his research yacht in Colombia, he interviewed planktonologists, oceanographers, statisticians and marine biologists in California, Alaska, and around Europe.

Robert published Dr. Dryden & The Missing Plankton at the height of the unedifying public name-calling.

The plankton-counting community welcomed the article as an even-handed explanation of both the scientific debate and the unscientific media ding-dong.

The mainstream media duly ignored it. 

They have enough trouble reporting how we’re trashing the air we breathe and the water we drink. The media don’t, it seems, have any appetite to add the collapse of our marine eco-system to the list of inconvenient climate truths. 

They may reckon the average reader finds things depressing enough already. They may be right. 

None of which changes the reality that the year-on-year PH balance of our oceans is rapidly approaching the point when it will dissolve the shells protecting the microorganisms that produce half our oxygen.

**

Among the tiny minority of people who care enough about this existential threat was a Dutch woman floating in the mid-Atlantic. She was one of the yachters who’d volunteered for Dr. Dryden’s citizen-science water-sample collection programme.

After reading the See Through News article, she subscribed to the See Through News free weekly newsletter.

Where, on Christmas Day, she read about Robert’s dilemma on how to find Global South candidates.

Bobbing on her Atlantic yacht, she thought of another newsletter she subscribed to, using up her precious internet bandwidth to read in her bunk.

She reckoned it might be worth putting Robert in touch with Renuka.

Dr. Renuka Thakore was founder of the Global Sustainable Futures Progress Through Partnerships Network. GSFN had a lot of similarities to See Through News.

Both were zero-budget social media networks.

Both were established around the same time.

Both had hit on the same trick. 

They were trying to use social media’s ‘free’ infrastructure for effective climate action. Our Silicon Valley Overlords provide free access to their platforms in exchange for our data, which they sell to advertisers, who use it to persuade us to buy more stuff we don’t need, pumping more carbon into the atmosphere. 

Robert and Renuka both had the idea of using this same infrastructure to reduce carbon by connecting like-minded people to take effective climate action. See Through News targeted a mass audience, while Renuka’s network consisted of…(and this was why our Dutch sailor, who prefers anonymity, made the connection)…sustainability researchers, mainly PhD and Masters students, mainly in the Global South. 

Renuka’s GSFN had grown, in a couple of years, to encompass 150 countries with thousands of researchers, all connected to a potential network of a million more. 

Maybe, thinks this Dutch sailor as she looks through her porthole into the mid-Atlantic Christmas darkness, maybe Renuka can help Robert find some needles in the Global South haystack in time to spend Tom’s half million dollars of supercomputing before the April 10th deadline.

She uses some more of her precious bandwidth to email Robert, and suggests he contacts Renuka.

This mystery Dutch connector didn’t realise, nor did Robert, nor Renuka, that they, Robert and Renuka were actually only a 90-minute drive away from each other in Southern England. 

There’s the power of social media, right there.

** 

On Boxing Day, before heading out for a walk, Robert sees the email from the Dutch sailor in his spam folder, takes a quick look at GSFN online, thinks it worth a shot, and pings Renuka an email.

In Episode 6, Rules, Referees & Global South Haystacks, we find out whether Renuka replies. 

Episode 6: Rules, Referees & Global South Haystacks

That Dutch sailor’s hunch was right.

Renuka and Robert hit it off immediately. Quite apart from the immediate matter of how to spend half a million dollars worth of supercomputer-grade cloud computing, they’re just immediately on each other’s wavelength.

Renuka and Robert have very different backgrounds, and until Covid lockdown, their careers have taken very different paths. 

Renuka, born in India into a family of engineers, becomes an engineer herself, before moving to the UK, into academia and consulting, focused on sustainability and carbon accounting, first in the construction industry, then everywhere.

Robert, born in London, studies Chinese, trades textiles for a Japanese multinational, views Asia as a producer and reporter for American TV news, makes documentaries on Europe for Japanese viewers, then, as the Internet destroys old ways of making a living, explores how filmmakers might use it to find a new one.

Yet by 2021 their paths start to converge. 

Locked down in Berkshire, Renuka founds the Global Sustainable Futures Network, a zero-budget social media network which, amongst other things, connects early career sustainability researchers in the Global South.

Locked down in Wiltshire, Robert founds See Through News, a zero-budget social media network with the goal of speeding up carbon drawdown by helping the inactive become active.

Both networks thrive and grow, and now intersect thanks to a good Samaritan bobbing on a yacht in the Mid Atlantic. One alert Dutch sailor spots that for The See Through Carbon Competition, Renuka’s network might be the answer to the problem faced by Robert’s. 

On December 27th, they have a Zoom call. Introductions over, goals aligned, with no need to consult anyone else or get approval for a budget, Robert and Renuka agree immediately to collaborate on the Competition, and start brainstorming. 

Before hanging up, they review their To-Do list:

  • Rules: must be clear and transparent, the simpler the better. 
  • Referees: assemble an Expert Panel to evaluate any entries, should they come.
  • Evaluation: define the criteria to pick winners
  • Support: cloud computing novices would need help to prepare their data
  • Marketing: advertise the Competition as widely as possible, while targeting the most likely candidates.

Even for problem-solvers like engineer Renuka and TV producer Robert, this is quite a to-do list.

Being optimistic do-ers, they don’t bother drawing up a similar list of obstacles, but it would look like this:

  1. The target Jan 1st launch date is in one week
  2. Most people in their networks are on holiday
  3. It seems no one has ever tried offering supercomputing to Global South researchers
  4. Most people don’t understand what the prize is 
  5. Experts are warning them suitable candidates might not even exist
  6. They also say even if they do exist, the April 10th deadline is nuts
  7. No one’s heard of See Through Carbon
  8. Even if any of this were possible with massive resources, they have no money

Daunting, certainly. Still, after seeing their zero-budget networks prosper for two years, Renuka and Robert have discovered something interesting about money. We’re usually all so focused on its advantages, we rarely consider its disadvantages. We think of money as lubricant, but it can also be sludge.

As former freelancers fighting for funding, Renuka and Robert are all too familiar with both of money’s isotopes, Renuka submitting countless applications for tiny puddles of grant funding, Robert pitching for endless meagre pots of documentary funding.

As GSFN and See Through News thrived without even a bank account to receive donations, the less obvious benefits of being money-free have revealed themselves. As many are quick to observe, you need to have the good fortune to have enough money to volunteer yourself, but many richer people keep working because they lack the concept of having ‘enough’ money to stop. 

Lubricant, yes, sludge, yes, but money is also narcotic. It’s a seductive means dressed up as an end. Running GSFN and See Through News has taught Renuka and Robert how much can be achieved by volunteers sharing an end goal that values carbon, rather than money.

What’s strange is how strange we find this. Money, after all, is an abstract, and very recent, invention of a particular species, homo sapiens. Carbon, by contrast, is what not only humans, but all life on earth, is made of. Put like this, valuing money so much more over carbon starts to look like the eccentric life choice.

Researchers and filmmakers can easily spend 90% of their waking, working hours pitching for money, waiting for money, accounting for money, or chasing money, leaving only 10% to actually research or make films. Remove money from the equation, and you’re instantly ten times more productive.

One thing Robert knew from his TV news days is there’s nothing like a deadline to focus the mind, and so it turns out when he and Renuka ask their networks for help. 

Even on holiday, GSFN and See Through News experts volunteer their help with alacrity, helping recruit others in turn to fill any gaps. 

  • Need the founder of an overstretched pro-bono PR agency on holidays? Happy to help.  
  • Experts to make the Panel as diverse and representative as possible? On board.
  • Agricultural expert to fill an expertise gap? Found one.
  • Graphic designer to create an eye-catching logo in a day? In your inbox.
  • A donor who creates solutions instead of problems? Meet Tom.

Renuka and Robert feel like they’re in the first half of The Magnificent Seven, The Dirty Dozen, or another heist movie where rag-tag teams of unlikely misfits with peculiar, but complementary skills, are being assembled on the hoof. Except on Zoom and with fewer shoot-outs.

On January 1st, they review the top of their to-do list:

  • Competition Rules- tick
  • Expert Panel of Referees – tick
  • Global Press releases to hundreds of global south publications for a broad appeal – tick
  • Academic call paper for a more focused appeal – tick

Unencumbered by the sludge of money, and with complementary networks of expert volunteers pitching in from around the world, with no need for anyone else’s approval, they hit Send.

On January 1st 2023, 4 months and 10 days before the deadline, The See Through Carbon Competition trumpets to a hungover world its quixotic search for golden needles in the massive Global South haystack.

In Episode 7, Golden Needles, we find out if anyone heard, and if the haystack contained any needles at all.

Episode 7: Golden Needles

Within days of its launch, it’s clear The See Through Carbon Competition faces a previously underestimated, but fundamental problem…

No one really understands what the prize is. 

As initial enquiries, and even applications, flow in from South Africa, Niger, The Philippines, and India, most are either saying they don’t know what’s on offer, or showing they don’t.

This makes sense. Robert isn’t doing a sustainability-related PhD, but when Tom rang him on December 15th about the YellowDog donation, he’d just spent months working on an in-depth article on AI’s dirty secret, and thanks to his unusual university flatmate, had a better-than-average understanding of computer science. 

Robert knew way more about the earthbound, carbon-guzzling reality of ‘cloud computing’ than almost anyone he met, yet when Tom told him about his donation of ‘ ‘half a million dollars of supercomputer-grade cloud computing’, he’d struggled to grasp exactly what this meant.

‘Can you explain it to me, he’d asked Tom, in terms of, I don’t know, multiples of laptops, GPUs, or’  – Robert was on very thin ice now – ‘multi-core cloud instances, or something?’, https://youtu.be/iBhmFxzYo1k

‘I could, and will’, Tom had replied. ‘but for the purposes of explaining this to potential applicants I strongly recommend you simply say ‘half a million dollars worth of supercomputer-grade cloud computing’.

Tom and Robert now collaborate on two Frequently Asked Question lists to publish on the website. One is a ‘Cloud Computing for Novices’ beginner’s guide, explaining key cloud computing terms and concepts, the other specifically for Competition applicants.

These are a big time saver for Competition applicants, but also makes most of them realise their projects don’t have anywhere near the volume of data, or complexity of processing, to require the prize on offer. 

One by one, they work out it would take them ages to work out how to prepare their data for supercomputing, and even if they did, it would barely require a dollar’s worth of compute. 

Robert was beginning to understand the vast gap between a decent laptop and supercomputing himself. He’d previously compared it to crawling with hands and feet tied together, then sprinting. He now realises it’s more like crawling, then driving a Ferrari.

All very educational for Robert and the applicants, but this only makes the prospect of finding any needles in the Global South haystack even more remote. The stream of enquiries slows to a trickle.

**

Then, a month in, one serious-looking application arrives from South Africa. 

Dr. Samuel Adekunle, a researcher at the University of Johannesburg collaborating with Manchester University in the UK. Samuel’s spent a couple of years collecting real-time data from air-quality monitors in and around Africa’s two busiest airports and city centres in Johannesburg, his adopted home, and Lagos, in his native Nigeria.

This is just the kind of potentially transformative project the Competition was designed for. Without access to supercomputing, Samuel and his team would have to make educated guesses as to what representative sample data to choose for processing and analysis. Access to Tom’s gift means they’d no longer have to guess, and could simply process the whole lot.

Moreover, Samuel’s application, and his response to the Expert Panel’s questions, indicate he and his team know one end of a supercomputer from the other.

Even better, Samuel’s ambition is to convert this heap of raw data into actionable, carbon-reducing research outcomes. 

In his video call with Renuka, Tom and Robert, as the Topic, Technical and Political experts on the judging panel, Samuel is impressive: strong technical grasp of how the prize could supercharge his research data, ticks all the See Through News boxes for helping the Inactive Become Active. 

Samuel is as impatient as the panel to rapidly convert results into on-the-ground measurable carbon reduction. Or as Samuel puts it, to ‘take lessons from the ivory towers to the common man in the streets’.

The panel approves Samuel’s application. Five editions after See Through News newsletter issue number 87 announced the launch of The See Through Carbon Competition, issue number 92 announces Samuel as its first winner. Everyone’s relieved. Not only do needles exist in the Global South haystack, it’s only taken a month to find a golden one. 

*

No sooner had the first needle been found, than another appears. 

This time it’s Laura Degiovanni, whose start-up uses blockchain and AI to sort out truth from lies for an app designed to promote trust and collaboration. 

Laura’s application wasn’t such a slam dunk as Samuel’s. As a London-based Italian, there’s no direct Global South connection. Also, it’s a commercial start-up rather than a research project. 

But Renuka, Robert, Tom and the rest of the Expert Panel agree Laura’s project ticks all the other See Through News boxes of accessibility, transparency, and open source universality, 

While not specifically targeted at the Global South, or involving Global South participants, Laura’s research would clearly benefit the Global South, maybe even disproportionately so. Poorer countries tend to have fewer defences against online disinformation, climate-related or not.

Because they’d had no idea whether any needles existed in the Global South research haystack, Robert and Renuka had drafted the rules to say the Competition favoured Global South applications, rather than being strictly limited to them. 

Not that they had to be accountable to anyone other than themselves, but they’re glad this decision allows them to consider Laura’s application, not least because there’s only a month before the deadline, when half a million dollars worth of supercomputing will go up in financial year-end smoke.

Another online interview with the Expert panel, and Robert drafts an article announcing another winner for Issue number 98 of the See Through News newsletter,

But before Sunday comes, there’s a third winner to add! 

**

When the March 19th newsletter comes out, 22 days before the April 10th deadline, it includes another name.

Johnson Jayeola is an industrial chemist and zero-carbon advocate who runs an eco-building and educational business based in Calabar, in Nigeria’s South-South state. 

His application appears absurdly overambitious. With less than a month before the computing deadline, he has no existing data and has designed a project in response to the Competition. 

Still, the scope of Johnson’s project – surveying firewood use for cookstoves in rural Nigeria, and sharing the data with national and international government bodies to promote its replacement with renewable energy, could hardly be more spot on. 

The Expert Panel may have its doubts about Johnson’s ability to have data ready from a standing start, but there’s no doubting his enthusiasm, passion, and grasp of the technical challenge.

Like Laura’s business, Johnson’s last-minute project doesn’t fit the expected profile of pre-existing PhD research, but Robert, Renuka and Tom decide that discovering such things, and being flexible enough to embrace them, is the point of doing something nobody’s yet tried. 

Besides, with 22 days before the deadline, why not take a punt? The YellowDog team say Samuel and Laura’s projects combined are unlikely to put much of a dent in the half million dollar prize fund. There’s little to lose by giving Johnson a chance. Off the record, Tom also mentions Samuel and Laura are struggling to prepare their data in time.

Then, the day after Newsletter 98 announced The Competition had a second and third winner, another last-minute entry becomes a fourth.

**

Dr. Daniel Zepeda Rivas, a Mexican researcher who’d studied, lectured and researched around Europe and the Global South, has spent years collecting two huge datasets from 500 locations around the world, predominantly the Global South.

One was a massive database of the carbon intensity of existing construction methods and materials in all these places. The other, equally massive, was the projected future weather and climate conditions for the corresponding locations. 

Use AI to process them together, and you’d end up with a huge, open source, online handbook that could tell anyone building a new building anywhere in the world how to do it the lowest carbon way, that would be most resilient to future weather extremes of wind, rain and heat. 

And building accounts for 39% of all carbon emissions. 

University College London and FCBStudios, a leading sustainable architectural practice, had provided Daniel with office space and support, but his project had stalled. It needed a massive amount of supercomputing horsepower to combine the two datasets. Maybe even half a million dollar’s worth…

With the clock ticking, and a massive amount of data to prepare, the Expert Panel fast-tracks Daniel’s application.

** 

On Sunday April 2nd 2023, the See Through News newsletter marks its 100th edition. It’s a doozy.

  • One article details the pilot for a new project, a free, open source citizen journalist project, called the Global Reporter Intensive Training Scheme, already training volunteers in the UK, Nigeria, Uganda and Kenya, aged from 10 to 45.
  • Another article explains how the global reach of the See Through News social media network has just shot past 70,000, is still growing exponentially at a double-digit monthly rate, on a marketing budget, including tax, of zero. (By the way, six months later it’s still growing at the same rate, and has passed 170,000. Same budget.)
  • And a third article, 9 days before the April 10th deadline, announces Daniel as the fourth winner of the inaugural See Through Carbon Competition.

By now, Robert, Renuka and Tom have come to see the Competition as a pilot, a proof of concept. 

Experts they’d consulted before launch doubted it was possible to find any needles in the Global South haystack.

They’d demonstrated they could. Maybe this would convince bigger, better-resourced donors, like those who actually owned the data centres, to make the See Through Carbon Competition bigger, more comprehensive and permanent.

In this sense, the fact that with no budget and a ridiculous deadline, the Competition had found Samuel, Laura, Johnson and Daniel, was job done. 

We now know that the Global South haystack not only contains such needles, but there are golden ones.

But with the April 10th deadline a week away, and all involved in the Competition experiencing the ‘groundrush’ sensation as an aircraft approaches the runway to land, a different question is looming larger.

Will any of the winners actually get their data ready in time to use the half million dollar prize?

In Episode 8, Tick Tock, we find out if any of them make it over the line.

Episode 8: Tick Tock

Days after announcing Daniel as the fourth winner of the See Through Carbon Competition, and with the deadline for processing counting down in days, Robert, Renuka and Tom get one piece of bad news, and then another.

First, Laura says she won’t be able to get her data ready in time for April 10th.

In the midst of a dotcom funding crisis, Laura’s team just can’t spare the resources to prepare their data for this unexpected deadline.

Perfectly understandable. Robert, Renuka and Tom wish Laura all the best in keeping her business afloat. They say if they manage to make the Competition permanent, her application is pre-approved for any future prize fund.

Then Samuel calls from Johannesburg. He’ll have to pull out too. He and his team face an academic version of the same problem facing Laura’s business. Transformative though the prospect of free supercomputing was, and helpful though the YellowDog team has been, Samuel has had to admit defeat. With no additional resources, getting their data ready by April 10th would be too much work, in too short a time.

Renuka’s network, Tom’s YellowDog onboarding team, and Robert’s website articles had done their best to help, but offering a free Ferrari has its drawbacks to people who’ve previously only aspired to a bicycle. You still need a driving instructor, road map and fuel.

**

On his See Through News name card, Robert gave himself the title of ‘Gift Horse Distributor’. It was a semi-joke to highlight the fact that See Through News had no bank account, and how some people only take things seriously if they involve money. 

The twin blows of Laura and Samuel pulling out brought the occupational hazards of the gift horse distribution business into sharp focus. Even gift horses need to be stabled, fed and groomed…

Neither withdrawal is entirely unexpected, both are completely understandable. What about the remaining projects? Will Johnson in Nigeria  and Daniel in London, the last-minute winners, get their projects over the line in time?

There’s also the matter of using up the half million dollars worth of computing. Proving winners not only existed, but could be found, was already a considerable achievement, but it would be a shame for that prize fund to disappear, unused, when YellowDog’s new financial year starts.

Samuel and Laura’s projects, despite their ambition, scale, and pre-existing databases, would still not have taken much of a chunk from the half million dollars. Tom reckons they wouldn’t have used up more than low tens of thousands of dollars worth.

But they were now out of the picture. 

**

With days to go, the only remaining candidates were Johnson’s, still collecting data, and Daniel’s, absolutely humongous and only just approved.

Robert checks in on Johnson in Calabar. Is there any realistic chance his data will be ready in time?

‘No problem’, replies Johnson. ‘All in hand and on schedule.’ 

Robert tries to think of Johnson’s project in terms more familiar to his own experience as a TV news producer. In 1997, as NBC Asia Shanghai Bureau Chief, he’d covered the Handover of Hong Kong to China at midnight on July 1st. 

Or a year later, now CNN Beijing Bureau Producer, he’d co-ordinated coverage of Bill Clinton’s official visit, the first by an American president since the first George Bush waved to TV cameras in Tiananmen Square in February 1989, weeks before the square was taken over by demonstrators. 

Both of these involved complex arrangements, in unpredictable environments, with constantly moving goalposts, but the hard deadline of a satellite booking window for live broadcasts. 

Not so different from Johnson’s challenge. Johnson had to design a project, obtain permissions from multiple government departments, create a data-collecting app, recruit dozens of volunteer surveyors, get them to the most remote parts of rural Nigeria, compile and consolidate the data, and prepare it for processing. In two weeks.

No problem.

**

How about Daniel? 

His project had the opposite problem, While Johnson was scrambling to  collect enough data ingredients from remote Nigerian villages, Daniel was calling in favours all over University College London to prepare his huge warehouse of data in time to squeeze through the cloud computing sausage machine.

Daniel was in daily contact with Tom’s YellowDog team, who had taken on the task with gusto. This was the kind of test engineers relish – a last-minute, epic technical challenge that would stretch their technology to its limits.

Deadlines are two-faced beasts that can both destroy and create. The April 10th deadline killed Samuel’s and Laura’s computes, but stimulated Johnson and Daniel’s projects into life, one from nothing, the other from a deep sleep.

Days pass. There’s nothing Robert and Renuka can do to help at this stage, so they stop asking how things are going. They can only distract Johnson and Daniel from their impending deadlines, looming larger with every hour.

***

On Sunday May 7th, nearly a month after the deadline, Robert hits Publish on See Through News newsletter issue number 105.

It contains a sneak peek of a documentary project about Britain’s greatest living woodsman, and a new episode of a podcast about a sumo crisis in 1993 Japan, both of which, being See Through News projects, were different starting points to the destination of measurably reducing carbon.

Also, a more obvious starting point, an article called  The See Through Carbon Competition Pilot Review.

It reveals what happened.

In Episode 9, Did It Work?, we find out if Johnson and Daniel got their projects over the line in time.

Episode 9: Did it work?

The actual compute is never going to be the problem for Johnson Jayeola. 

Before he applied, he read the cloud computing Frequently Asked Questions and gave himself a crash course in what was on offer. 

Once approved, before he starts collecting any data, Tom’s YellowDog onboarding team explains how to structure it so it’s ready to go. 

Johnson and YellowDog both know it won’t use up much of the half million dollars. Even with Johnson’s optimism and ambition, there’s only so much data he can collect in a few days.

For Johnson, it’s all about meeting the deadline in a country where few things happen quickly, with an abundant supply of unexpected problems, from palms that need to be greased to terrible roads and poor telecommunications.

While his team develop and test the app, Johnson works ministries and government offices. In formal letters, official phone calls and informal meetings, he explains the importance and urgency of his rural firewood survey to potential volunteer partners. 

By now, Johnson and Robert are chatting quite regularly. As well as discovering Johnson, in coastal Nigeria, is just as big an Arsenal fan as Robert, who grew up near their North London stadium, they enjoy chewing the fat on a variety of topics.

Johnson tells Robert how he comes from a village with the highest concentration of PhDs in Nigeria, and as the youngest of four sons, he’s his father’s last chance of joining the ‘My Son The Doctor’ club.  

‘What you have to understand, Robert’, Johnson explains on one video call, ‘is that the biggest occupation of any African man is boasting’.

Hearing Johnson’s methodical efforts to get the government offices, volunteers, coders etc. on board, Robert tells Johnson it reminds him of a key lesson from his first job, at a Japanese global trading company. ‘What you’re doing’  he tells Johnson, ‘sounds like an African version of nemawashi’, 

Nemawashi, literally a horticultural term meaning ‘tending the roots’, is a key part of Japanese corporate culture, and the reason why formal meetings are so short in Japan. 

They only have meetings, and formal votes, when they know everyone is on board, having spent the previous weeks and months tending roots in one-to-one meetings to get everyone aligned and invested in the project, usually in karaoke bars in the small hours.

But Johnson only has days, not weeks or months.

Young volunteers are keen to help with the legwork, but like young people everywhere, young Nigerians can’t afford to pay for travel, food, accommodation, phone and internet costs.

Johnson puts his hand in his own pocket to fund the incidental costs, but otherwise relies on resourcefulness and persuasion. 

Working his personal, church and professional networks, he assembles his team.

It’s a Nollywood version of the The Magnificent Seven re-enactment that Robert and Renuka did over the Christmas holidays a few weeks previously, converting Tom’s donation into the See Through Carbon Competition.

Johnson sends letters to key people, leverages press coverage of his win, hits the phones to find volunteers to do the considerable legwork. 

The days click over towards April 10th.

Johnson plugs into Ploggers Nigeria, a network of young Nigerians who volunteer to combine exercise with environmental good deeds. This usually involves jogging around Nigerian streets picking up plastic trash, but Johnson convinces them to extend their mission. 

He invites Robert to dial in to an online training session for the Ploggers. 

  • You’ll be travelling to Nigeria’s poorest and most rural regions, he tells them. 
  • Use this bespoke app we’ve just made to survey locals about their cooking methods. 
  • Make sure you can speak the right language, figuratively and literally. 
  • And while you’re at it, educate the villagers about affordable low or zero-carbon alternatives fuelled by wood pellets or solar energy.

If cooking with firewood seems like a ‘yesterday problem’, welcome to the Global South. While the Global North frets about disposable coffee cups and plastic stirrers, around two billion people – including those in relatively rich countries like Nigeria – cook with firewood every day. 

When the experts at Project Drawdown rank dozens of human activities for their carbon-reducing potential using current, proven technology, cookstove technology is number 9.

The short-term public health consequences of cooking on open fires are just as disastrous as the long-term environmental ones. Still, for such a huge problem, there’s been very little systematic and detailed research –  maybe because the people funding such research think of it as a ‘yesterday problem’. 

Now, armed with his just-developed data collection app, Johnson’s team of volunteer Ploggers are uploading survey data from villages around Nigeria. 

Somehow, Johnson collates the data and sends it to his new friends at YellowDog, who complete the compute with a day to spare.

Quite apart from any contribution to speeding up carbon drawdown, this is a bravura logistical achievement. 

Congratulating Johnson on a Zoom call, Robert tells him a story about his father, an eminent neurologist. One of his patients was the boss of a famous footwear brand, who at the end of a consultation gave Robert’s father the highest compliment he knew. ‘Dr. Stern’, he said, ‘You would have made an excellent shoe salesman’.

‘You’, Robert tells Johnson, ‘would make an excellent TV producer’.

**

While Johnson is conjuring up his data from scratch, the YellowDog team is knee-deep in a very different challenge – how to squeeze Daniel Zepeda Rivas’s huge inventory of data through the cloud computing sausage factory before time runs out.

To satisfy the Competition’s Global South remit, and to make the compute manageable within the time constraints, Daniel thinks it’s prudent to select only part of his enormous database. 

Conferring with Robert, he selects his native Mexico, and East and West Africa from the 500 locations available.

On April 1st, Daniel and his University College London team primp his two datasets from these Global South locations into compatible shape, and place them in YellowDog’s in-tray. It is now a simple race against time to process it all by April 10th.

Given that this whole adventure had started with Robert interrogating YellowDog’s low-carbon claims, for his article on AI’s Dirty Secret, Tom’s team do their utmost to minimise the carbon footprint of Daniel’s compute.

For the first few days, they survey the tens of thousands of data centre options they constantly monitor, and only select those with the lowest carbon energy source. 

This varies from month to month, but in early April, the lowest-carbon options available are in Paris, powered by nuclear energy.

It goes well. Daniel and YellowDog get more ambitious. Maybe it’s possible to prepare and process data from all 500 locations before the deadline… 

At this stage, everyone involved is so focused on the technical challenge, no one mentions whether this might blow the half million dollar budget.

**

Day after day, Daniel and his team dress more of their data and deliver it to YellowDog, who distribute the task to data centres around the world to subject it to AI analysis before April 10th.

As the deadline approaches, they realise they have to be less picky about the carbon credentials of where they allocate the jobs. 

They measure the carbon intensity of data centres by grammes of CO2 equivalent per kilowatt hour. 

Paris comes in at 29, but they now have to use data centres in Milan that are more than 7 times more carbon intensive. 

Then, as the tick-tock grows louder, they add London and Frankfurt, 8 and a half times dirtier. 

Finally, as midnight on April 10th approaches, it’s whatever’s available. They complete the compute in data centres in Phoenix, Arizona, where two-thirds of power is from coal and oil. 

That makes the Phoenix compute 24 times more carbon-intensive than the nuclear-energy-powered Paris data centres they started with.

But complete it they do. Over 11 days, they cram in 45,000 supercomputer hours.  If it means anything to you, the output from all that supercomputer-grade number-crunching was 36 Terabytes. If not, it’s a lot. 

Johnson’s project may only have driven the Ferrari a few feet, but Daniel’s went on an epic joy ride with the accelerator jammed to the floor.

**

A few days later, YellowDog does the carbon accounting. It’s fascinating. Robert updates his article on AI’s Dirty Secret. For details, see the See Through News website and scroll to the end of the article.

Robert does his best to translate the scale of Daniel’s compute from terabytes and instance hours and iterations into more easily visualised images.

Doing some back-of-the-envelope calculations as he prepares the Sunday newsletter, Robert decides AI’s Dirty Secret might be better measured in charismatic macrofauna than grammes of CO2 equivalent per kilowatt hour.

Had there been no deadline, he writes, Daniel’s compute could have used only the lowest-carbon data centres, the ones in Paris. These would have only emitted 541 kilogrammes of carbon into the atmosphere. About the weight of one Right whale testicle.

Because YellowDog had to resort to dirtier data centres to get the job done in time, it ended up emitting 5.8 tonnes. About the weight of an adult elephant.

Had they done the entire compute at the dirtiest data centre, in Phoenix, it would have emitted more than 13 tonnes of carbon. About the weight of an entire female sperm whale.

Does that strike you as a lot? If so, what are your thoughts on the carbon cost of all those streaming services, video calls and AI apps that produce pictures of the Pope in a funny hat?

Or, for that matter, the buildings we live in. In 2022, the global built environment pumped out four billion tonnes of CO2 equivalent into the atmosphere, through construction, usage, repair and maintenance. That’s 4 followed by 9 zeroes.

Or the testicles of four billion right whales.

Or, if you can imagine a fully crewed, loaded, armed US Navy Nimitz Class aircraft carrier, around forty thousand of them.

And what about the prize fund? How much of the half million dollars did Johnson and Daniel’s projects end up using? 

The answer, which might surprise you, can be found at the end of the revised AI’s Dirty Secret article

If you want to know the answer right now, remember the second part of the See Through News Goal – ‘Speeding up Carbon Drawdown by Helping Make the Inactive Active’.

If you’re really curious, finding and reading the article only takes a few seconds, but it does require a tiny action.

But more to the point, was any of the effort described in these nine Episodes actually worth it?

In our final Episode 10, What Happened Next?, we’ll let you be the judge.

Episode 10: What Happened Next?

It’s September 10th 2023, 5 months after the deadline for the inaugural See Through Carbon Competition. 

The 10th being a Sunday, it’s See Through News newsletter day.

Robert’s writing an update. Not much work today – he’s just editing the summaries Johnson Jayeola and Daniel Zepeda Rivas have sent him, and adding a few comments of his own.

Johnson’s update details the logistical complexities that would make him such a great TV news producer.

His research collected and collated data in real time, about 5 cooking instruments – firewood plus 4 comparators, in 12 States across Nigeria, involving 20 Local Governments and 36 Communities, mainly in the South West.

Output from the cloud computing analysis include granular breakdowns by cooking instruments and geography, and data visualisations.

Johnson describes a unique, precise, accurate dataset, made possible by the See Through Carbon Competition, that’s already having real-world impact on the ground.

Government agencies, aid organisations and NGOs now have statistically rigorous evidence to guide informed decision making on maximising carbon reduction through the targeted provision of clean-energy cookstoves.

In keeping with See Through New’s principles of transparency, Johnson has made his data freely available, and his methodology open source, to encourage similar research across Africa and beyond.

**

How about Daniel? What has he done with his 36 terabytes of output?

For a start he’s had to scramble to find anywhere to store it. 

Even these days, 36 Terabytes is a serious volume of ones and zeroes. 

Daniel persuades University College London to hold it while he works out backups. He spends weeks transferring the data to a couple of stacks of hard drives, with multiple redundancy for insurance. 

The stacks humming on his desk are about the size of a cardboard box. They feel as hot as a data centre processor.

Meanwhile, he and other architectural and engineering experts have been spot-checking the output. As we now know from ChatGPT, trained on the Internet, Robots are only as smart as the data they’re trained on. 

There’s always the risk that despite all Daniel’s efforts in collecting, matching and preparing the data inputs, the AI processing had generated nonsense outputs.

Their spot-checks check out. They all make sound engineering sense. Particularly interesting are some results that initially look wrong, weird or counter-intuitive, like only insulating certain faces of certain buildings with certain usage and occupancy. 

No architect or builder would ever consider anything other than symmetrical insulation – only insulating, say, the north and east faces of a building just looked wrong. But when they ran the numbers, there are indeed certain conditions where selective insulation would reduce carbon both in construction and ongoing heating costs. 

This is just the kind of thing robots are good at – spotting patterns that elude even human experts, used to seeing the world in a particular way. 

It’s a cracking example of how judicious use of a precious carbon-intensive resource – in this case AI processing – can be justified because it reduces vastly more carbon in a way no number of human experts would have worked out. 

Unlike, say, spending hours on video calls, streaming movies, or generating pictures of the Pope in a funny hat.

With his colleagues at UCL and FCBStudios architectural practice, Daniel is doing some building himself. 

They already have a working demo of a tool to help anyone building a new building, or retrofitting an old one, to save the maximum amount of carbon while building the most robust construction.

Whether you’re building a hut in Namibia, or a tower block in Toronto, the design tool will consult the handbook made of those 36 terabytes of computations, and tell you the lowest-carbon way to make your structure most resilient to future climate extremes.

Anyone building anything anywhere in the world will be able to consult Daniel’s free online database, to construct something that will withstand extremes of heat, flood and wind, when traditional building practices and methods no longer work.

And Daniel’s handbook will mean the new way they design and build will minimise putting more carbon in the atmosphere, and making future extremes of heat, flood and wind even more extreme.

In his summary, Daniel makes clear what has made this possible. He writes 

the half million dollars allocated to the project was fully utilised the best way possible. This additional funding allowed us to overcome the technological challenges we were facing and expedite the process, resulting in a more robust and comprehensive analysis.’

Daniel and his colleagues are on track to make their handbook, and design tool, publicly available to anyone on the planet in a matter of months.

It has the potential to massively reduce huge amounts of carbon.

And that was the only condition Tom placed when he rang Robert to make YellowDog’s donation, 9 months before. 

There you have it. A neat ending.

But is this story over? Will there be further episodes? Don’t know.

Was Tom’s unusual gift a one-off, or the start of something huge that could measurably speed up carbon reduction, particularly in the Global South, for the foreseeable future?  Don’t know.

Watch this space. Maybe you can tell us. Maybe you can take action to change the story.

This is a real-life, real-world cliff-hanger to which no one knows the answers.

Here are a few unanswered questions.

  • Will Johnson’s open-source methodology and data convince governments in Nigeria, and other African countries, to act on firewood use for daily cooking?
  • Who will use Daniel’s free online low-carbon building design tool, and how much carbon will it save? 
  • Will YellowDog’s low-carbon approach convince investors it deserves to survive the venture capital drought now ravaging hi-tech startups?
  • Will Renuka’s global network continue to grow and foster researchers in poor countries without any funding or formal support?
  • Can Robert leverage the Competition to benefit See Through Carbon, the radical new carbon auditing standard launched Sept 2023?
  • Will we continue to treat cloud computing and AI as an infinite resource, and ignore carbon consequences?
  • Is the story we’ve just told you a one-off? Or can it convince big cloud computing players to donate processing, mentoring and expertise to make the Competition bigger, more comprehensive and permanent?
  • Or is this the end of the story?

[sting]

AI’s Dirty Secret, or How To Spend Half A Million Dollars of Supercomputing was narrated by George Hinchliffe, who also wrote the music.

It was written and produced by Robert Stern.

The See Through News Podcast is a See Through News production.

See Through News is a zero-budget social network with the goal of Speeding Up Carbon Drawdown by Helping the Inactive Become Active.

For more, visit seethroughnews.org

Thank you for listening