Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage Presentations Climate Change & Tech, Good and Bad Bedfellows

Climate Change & Tech, Good and Bad Bedfellows



Paul Johnston explores the impact of climate change, what kind of impact technology companies are actually having, some strategy tools to use, looks at the changing world of business continuity and disaster recovery and shows how to identify high and low impact interventions.


Paul Johnston is an interim CTO and strategist who has particular interests in serverless, cloud, startups and climate change. Formerly, he was Senior Developer Advocate at AWS for Serverless and CTO of multiple startups including one of the world’s first serverless startups.

About the conference

Software is changing the world. QCon empowers software development by facilitating the spread of knowledge and innovation in the developer community. A practitioner-driven conference, QCon is designed for technical team leads, architects, engineering directors, and project managers who influence innovation in their teams.


Johnston: We talk about whether or not climate change and tech are good things together. Often, we think that tech is good for everything, when we talk about it in the tech world. We like to think we're the good people. I'm going to try and basically explain to you that we're not. Or, at least I don't think we are, and certainly, not all of the time.

Let's talk about climate change. Let's talk about Hurricane Dorian which hit the Bahamas last year, and had 185 mile an hour winds. It was the worst storm, strongest Atlantic storm to ever hit, certainly, that we know about. Seventy thousand left homeless, $8 billion in damage, huge storm hitting the Bahamas.

You've got Typhoon Hagibis, which hit Japan. It had 140 mile an hour winds. It had a meter of rain in 24 hours. Just imagine thinking about that. Six million people affected. One of the things I found out was that the typhoon season in Japan and that area of the world is getting longer due to climate change.

Then you've got Cyclone Idai, which a lot of people have never even heard of, which hit Southeast Africa, which hit those 4 countries. Three million people were affected, 2000 people missing, 1300 people were dead, $2 billion in damage. The reason there's less damage is because there's less infrastructure. It's just a poorer country. Most people have never even heard of this. All those happened in 2019. I think we like to talk about the things that happen in rich countries and to places that matter where the media has lots of people and outlets that we can talk about.

We also like to talk about Storms Ciara and Dennis. Whatever happens here in the UK is really important to the people in the UK. Four hundred million pounds of damage happened in those two. We've got Storm Jorge happening right about now, or was supposed to be happening. It hasn't really happened. Something actually very important happened, which was a category 5 atmospheric river. Basically, a huge amount of water went up into the atmosphere and got dumped on the UK. Category 5 is basically classed as exceptional. Exceptional is it doesn't happen very often. Then we had a bomb cyclone, which is not very often something like that happens. They happened a week apart. They were essentially exceptional events a week apart.

Then you've got Subtropical Storm Kurumi, which happened in January 2020 in Brazil. There was the highest daily rainfall in Belo Horizonte in 110 years. You've got all of these events, and flooding, landslides, and evacuations, and all of those things. They happened in 2 months of 2020.

Where Is Global Warming Going?

We like to say actually, of course, storms have always happened. We know storms always happen. There have always been bad storms. What we know is that climate change is changing the way that these storms happen. They're getting worse. They're not getting more. That's the point that we're getting. You get stronger storms. You get slower storms. You get larger storms. You get more rainfall. That's what we're seeing. One of the things that people like to say is, "It's actually not getting hotter." It isn't actually getting hotter if you just take the atmospheric temperature. What's actually happening is all the heat from global warming is going into the oceans. That will actually change over time, because the oceans can only absorb a certain amount of heat from the sun. Then the atmosphere will start to warm up. That's what we're beginning to see now. That was from 2003, that graphic.

What we're actually seeing now is the beginning of the heating up of the atmosphere as well. What that leads to is droughts. Droughts are now beginning to happen. We're beginning to see rising global temperatures. Those droughts then lead to different conditions. Something that came out overnight was that the summers in Australia are now a month longer than they used to be 50 years ago. You've got all of this change that's happening. Climate change isn't just the sea levels are going to rise, and we're going to lose all the polar bears. We've got all of these changing patterns of climate over long periods that are making things different. If someone throws a snowball at you and says global warming isn't real. Don't believe them. It's a complete load of rubbish. We're getting all of these wildfires that are precipitated. We've had wildfires. We have wildfires. They get worse because of climate change. Actually, the wildfires, no one predicted the level of wildfires that happened in Australia because of climate change. They were actually worse than was predicted by climate change.

Then you get the extremes, which is also that you got the heaviest rainfall in Sydney for 30 years, straight after you had these terrible wildfires. You get all of these events happening. You get the extreme weather, which becomes the new normal. Then you get all of this happening all together. You've got all of these events. You've got climate change. You've got all of this.

Then you get me standing on a stage talking to you about all of this. You come to a tech conference, the first thing you come to, standing here, why am I talking to you about this? I'm a technologist. I've basically been in tech for over 20 years. I started off with servers in a small data center in a university in the UK, worked my way up to being a CTO. I've worked at AWS. I also am a climate activist. I am a pro bono director of a community energy company. I've spent a lot of the last two years spending my time talking about climate change, the intersection of climate change and technology. I've had a lot of conversations with people about how climate change and technology mesh together and what people should be doing. A lot of the time, it's I want to create an app that helps people to figure out their carbon footprint. Basically, if that's your thought, please go and read this blog post that I wrote because that will basically stop you doing that and help you to understand what you should be doing. If you're not sure what to do about all of this, and get a bit of an idea of where to go, that's a good place to start.

Let's get back to how this works, what we're talking about. How the extreme weather, and climate change, and all of those other knock-on effects may actually start to affect tech too, because they do. Hurricane Sandy which was a massive storm in 2012 actually took out 8 data centers. I don't think most people know that. Most people just think, "It was just a massive storm." It took out 8 data centers. There was a basement data center in New York that got flooded. It was recently built. Because it was below ground level, and all the flooding happened at surface level, the flood actually went into the data center. Flooding-electricity, don't really mix. It's a bad idea. There was another one where the diesel pumps that kept it going, because the electricity had gone out, they couldn't refill the diesel. That's what you do when you have a data center, you have diesel backup. They couldn't get the diesel to the data center because of the floods. The diesel generators ran out and it ran out of electricity. It's not just that you've got all these floods, it's the secondary impacts. Then you've got other issues.

There are other places that actually had to deal with this as well. O'Reilly is another organization that had to deal with this. Fascinating one, power outages in California because of wildfires, what happened? Building was in an evacuation zone. They had their data center in their head office building. They had to evacuate. They couldn't get back to fix something, weren't allowed back in the building. Power went off. What happens next? Everything runs out. You can't go back to fix it. Everything goes down. One more reason to be in the cloud, it's all tough.

Let's just give you a little bit of background, so where we're going. What's going on? How the world is reacting to all of this. Because I think most people have an idea that actually people are dealing with this and it's ok. Let's have a look at the IPCC. The IPCC was set up at the United Nations. It's the Intergovernmental Panel on Climate Change. Intergovernmental, lots of governments talking to each other. There are 195 member governments, including countries that don't really want climate change to be happening, but lots of them. They put together a panel of people. If you imagine, basically people. They appoint. They basically then assess all of the climate change research. They put out reports. It's quite important that you think about that. It's basically all of the governments of the world saying, this is the research, and this is what we say. An IPCC report comes out, you can pretty much say it's an authoritative report put out by the governments of the world, pretty much.

Paris Agreement, 2015, all of the governments of the world came together and went, the aim is to keep the global temperature rise by 2100 to well below 2 degrees above pre-industrial temperatures and pursue efforts to limit to 1.5 degrees. That was what they all said. That was the plan anyway. One of the things that the IPCC did was go, let's try and explain to everyone what would happen if we got to 1.5 degrees.

They created a special report, which came out in October 2018. If you have not read the special report, or even just the headlines from it, I recommend you go and do it, because it will scare you. It should scare you. You should be worried about it, because it includes a whole bunch of pathways for understanding what that means. It's where the idea of the 12 years comes from. We've got 12 years to save the planet. We don't have 12 years to save the planet. We just have 12 years to try not to overshoot 1.5. That's pretty much what we're talking about. The point about it being scary was real. This is from the IPCC, which if you remember, is 195 member governments actually putting all of this together. We're not talking about a whole bunch of scary people trying to scare the rest of the world. We're talking about 195 member governments putting out a report. We're not trying to say, it doesn't really matter. It's actually quite a big deal.

How Are We Doing?

That was October 2018. How are we doing? The UN has updated itself, last month. We're on course for 3.2 degrees. We're getting nowhere near. I think it's quite important to update you on all of this, so that we can have a bit of context for the rest of the conversation and talk about technology a little bit. What they basically said in February was we need to drop global emissions by 7.6% a year for the next 10 years to hit 1.5 degrees. Last year, emissions went up. We're getting nowhere near. That's why people were like, why I'd say it's time to say goodbye to our most ambitious climate target ever. Basically, the governments of the world aren't doing anything. What they also said in their release was that every city, region, business, and individual needs to act now. We all need to be a part of this. We all need to be thinking about this. We all need to be doing our bit, which is why I'm talking to you about it now and why I'm mentioning it now. It's not about someone else is going to do something. We all have to play our part, whatever that part is. Play a part in some way.

The interesting part about this is that it's not just me saying this. Some of the biggest companies in the world such as AXA are saying, if we reach 4 degrees, which is 0.8 degrees above 3.2, and 3.2 is average. There are aerobars within that. 4 degrees is an uninsurable world. If we get to an uninsurable world, in the context of massive companies like AXA, who insure the world, I'm pretty sure you can see that that leads to some consequences for the economies of the world that are pretty much not ones that we want to see.

Then you end up with organizations like BlackRock, massive investors, huge investors going, we're not going to invest in fossil fuels anymore. This is where we are. We're going into a world where the pressure is now hitting, where companies cannot actually say, we don't need to worry about this anymore. Actually, companies need to worry about this. Your organization, whoever you work for, needs to worry about this. They need to start worrying about it now because climate change is already here. It is going to change our way of life in some way. How it's going to change it, I do not know. It is going to change your working patterns. It's going to change what your business does. It's going to change every part of that in some way. How, I'm not sure. We will get there.

For many, it's already changed. Mainly, those people are in the developing world. Actually there are reports that say that the impacts of climate change are going to be primarily on the developing world by about two-thirds to a third. When we talk about impacts, people in the industrialized world, the developed world are actually only going to see about half as many impacts as the developing world. Actually, what's going to happen is that our lives are going to change less than those in the developing world, which means that actually we don't see the effects as badly. I don't think that's fair. That's a whole other conversation.

What Has This Got To Do With Tech?

What's this got to do with technology? Tech has started to get all climate-y. I put it in quotes for a very good reason, because we started to see things like this. People going, I'm going on a climate journey, and starting a podcast. Then we're seeing the CEO of GitHub turning around and going, "If you're using GitHub for climate stuff I'd like to know." I'm not entirely certain what that means. It's like, what are you doing with it? How are you addressing GitHub with tech? I'd love to know what he means by that. Then we're seeing Sequoia, saying, "We'd love to invest in climate tech and sustainability." Exactly what does that mean as well? You're just looking for more ways of making a lot more money. Or you're actually trying to look at the systemic problems within society. What are you trying to do? You're just trying to make money, or you're actually trying to address climate. There are issues as to what that actually means.

The question I come up with is, are we on the right side of this one? Because I think there's a real issue here. Technology often comes across as very positive. We are people that spend our time going, "We're doing the right thing." We talk about ourselves as doing the right thing all of the time. We think we're on the right side. We think that the stories are always positive. We always come up with AI is going to change the world. Machine learning is going to change the world. Everything is going to change the world. We are the ones. We are the kingmakers. It's always the story that I see. We never ask ourselves, are we the baddies? We never do that. We never actually sit back and say, are we the ones on the wrong side of this one? We never sit back and ask ourselves that one.

Part of that is a quote from Hannah Arendt. I think this is something that tech people should think about an awful lot more, which is the sad truth is that most evil is done by people who never make up their minds to be good or evil. I think if you imagine someone like Mark Zuckerberg, I think this quote almost sums him up. He's sitting there going, "I'm not going to make any decisions. It's up to the public. It's up to the people that use our platform to work out what they want to do." He's almost the epitome of this quote. We take it. We lap it up as technologists.

You can see that. If you have a look at the Stack Overflow developer survey from 2018, they put in some ethics questions. I couldn't see them in the 2019 survey, just so you know. Don't know why they took them out. Over 80% of people who answered, who's ultimately responsible for code that can accomplish something unethical? Eighty percent of people said it wasn't the developer that wrote it, someone else's responsibility, someone else over there. It's not your responsibility. It's not mine. Even though I wrote it, it's not my responsibility. We deflect.

Often, the way that that's put forward is so often utilitarian in ethics. Happiness and well-being, I just want to make my money. I want to make sure that I've got a good life for my kids and my family. I'm ok. It doesn't really matter. Often we hear the words unintended consequences when we talk about AI. Unintended consequences, we didn't mean for that to happen. Often, that comes out in tech. We built this. We didn't intend for that to be the case. I asked on Twitter the other day, what would be the best way of explaining utilitarianism? My friend John Nolan came up with, "Kill one to harvest organs for five." That's pretty much how tech thinks. Genuinely is, if you think about it, you ask a tech person in a slightly less controversial way how you would do something and they would think like this. I'm not saying you would. If you ask the broad population of tech in a slightly less obvious way, they would come up with an answer like this. That would be my thinking about how I've seen the tech world.

The shift towards employee activism at places like Google and Microsoft, and Amazon, around climate change and tech is a significant shift, in my view. Because we've seen employees of big companies join Climate Strike. They've done more than that. Here's a typical banner, which is the No AWS For Oil and Gas. If you don't know, all the major cloud providers, as far as I'm aware, are involved in helping oil and gas companies extract more oil and gas, which is slightly strange. There are some ethical problems there.

When we talk about tech, we talk about data centers. We're talking about data centers also producing about as much greenhouse gases as aviation and shipping, in the same ballpark. We're talking 1% to 2%, something around that ballpark. There was a piece of research that came out last week that said that we haven't actually grown that much in the last year. The interesting bit is the forward-looking research that looks like it might grow about three times. Right now, we haven't grown that much over the last 10 years, but probably going to grow an awful lot over the next few years, mainly because of machine learning and AI, and all the data stuff that we're going to be putting into the data centers over the next few years. Aviation, pretty bad, if you fly backwards, Heathrow to JFK and back, it's about 3 tons of CO2 equivalent per person. Cloud is actually more efficient than data centers, pretty much stands to reason. I'm just telling you that.

The thing is, Jevons Paradox. If you've never come across Jevons Paradox, go and look it up. Efficiency, if you make something more efficient, you use more of it, which is more straightforward. If you imagine, you've got a data center. If you want to put something into a data center, you have to buy a server. You shove it in. Then you will be more careful with the resource. If you're using virtual instances, you can spin up an instance in a few seconds. You'll use a lot more of them, because you don't have to think very much more. Actually, having cloud that isn't powered by sustainable electricity is a problem.

Just as a quick example, on Jevons Paradox, if we make machine learning models, and AI, and all of this, if we make it easier to create the models, we will create more of them. If you make it simple to create the models, everyone will go and do it, which is what we're seeing now. We're seeing an explosion in the amount of people who are using AI and machine learning to do stuff. When it was hard, when you had to have specialist hardware, not many people did it. Now that you can do it, now that you can go and create services on any one of the cloud providers, actually a lot more people are doing it.

Quick aside on the ethics, this came out yesterday, from "The New York Times." China is using AI to regulate citizens to limit their movement due to the Coronavirus. This is what people do. You basically can't walk through certain checkpoints without your app being green. If it's red, you get blocked. This is what people do with tech. If you want to learn about efficiency in data centers, there's a brilliant Jessie Frazelle blog post. Go and read it from last Wednesday.

How Good Are The Major Cloud Providers At Being Green?

Let's talk about data centers, and data, and how green they are. Because I think it's really important. Good news, Google and Azure are pretty sustainable. AWS is not. There was a data center ethics white paper that myself and my friend Anne Currie wrote in 2018. It had an update last week. I'm about to show you the new updated figures. Basically, Google and Azure come out very well, for lots of very good reasons because they are trying to do the right things. Google comes out higher because they try to do hourly based time matching of renewables with their usage. If they use renewables for an hour, they will try and buy renewables for an hour in certain data centers. In places like China they don't do it very well. Azure, they are trying to go carbon negative, of Microsoft's entire history by 2030, a whole other conversation. AWS, you can't get a carbon footprint for your own systems, and only four green regions. It's not good. From a green point of view, it's not very good.

Microsoft, they're trying to go carbon negative by 2030. Basically, they're hoping that someone's going to create some technology with this billion dollar carbon fund. Then Jeff over here has said he's going to pledge $10 billion for an Earth Fund, which is really fascinating, because over 400 of his employees are risking their jobs to call out Amazon on their climate crisis failures. I don't understand how that works. He's putting a whole bunch of money but isn't allowing his employees to criticize. It just is a bit strange.

AWS and all the cloud providers are still funding oil and gas extraction conferences, and things like that. It doesn't make any sense. I've spoken to people in the developing countries. They are basically completely confused as to how we should be working around all of this. They don't want the tech companies to do this. They still are. Amazon when they released their climate pledge in September last year, they released data about their emissions. They're in the top 200 emitters of greenhouse gas worldwide. Top 200, they're not one of the good guys at the minute. Are they greenwashing us? Possibly, maybe we should as customers be telling them so.

Tech Has Some Real Problems It Needs to Solve

Tech has some actual real problems it needs to solve. We're not in this to be saying that just because you're in tech, you're good. You can't just be saying, it's someone else's problem. People talk about offsets and planting a trillion trees. That will be fine. Let's all make the bad climate stuff magically disappear. Offsets are not going to solve it here unless we can do it at scale. We just don't have the scale to stop the consumption. Trees are only going to get us so far. We need a whole bunch of other stuff. Just offsetting is only a simple thing.

When the IPCC did their reports in the 2018 report, what they basically said was, we have no scale to take carbon out of the atmosphere. We have a carbon budget. Offsets are basically you emit and you stop emitting somewhere else, but you're still emitting. You still push carbon out. You have to reduce emissions by a lot, basically, to take carbon out. We almost certainly have to take carbon of the atmosphere.

Basically, it's a brilliant illustration. If we'd have started 15 years ago, the curve is relatively flat. It just goes down. It's quite nice. The further we go, the steeper it gets. Limiting warming to 1.5°C is increasingly difficult without large scale negative emissions. This is zero. If we don't have negative emissions, we're not going to be able to do it. We don't at the moment. I'm trying to put this into tech words. Offsets are a leaky planetary patch while we try to come up with a permanent fix. We don't have one yet. When someone tells you someone's going to create fusion power. Fusion power has been 35 years away for about 50 years. It's not coming any time soon. Please do not expect that to happen. In fact, pretty much every technology isn't coming any time soon that I can see.

What Can I do?

Let's talk about what you can do. We got extreme weather. We've got big companies being green. We've got all of this carbon budget stuff. We've got cloud. Told you what I think about what you should be doing. Read my blog post. Let's talk about something quite simple. Because it informs how we should do tech. I think this is probably the most important part of the talk. If you go away with nothing else, this bit. This slide may be the next one, because it's got an extra bit on it. Most people when they see reduce, reuse, recycle, just remember the bottom bit. They just remember the recycle bit because that's the bit everyone knows about. Everyone thinks about these three words being about recycling. It's not. The most important bit is the top one. It always was the top one. It comes out first. Everyone remembers the last one.

Actually, I would add in another part. It's called the waste hierarchy for a reason. If you don't create the waste in the first place, it's even better. One of the things we don't talk about in tech is not building stuff. The best climate solution is not to build it in the first place. It's quite important actually, to really think about, do we even need this? Because I have serious concerns that actually most tech people spend their entire time going, "This is really clever. Let's build it." Without going, "Do we need it," as the first question? One of the problems I have is that most people don't have enough tools to even start to have that conversation. Let's talk about one tool. In terms of all of these things, think about this in terms of how you do your tech. Think about it in terms of how you run your projects. Start there, and then go from there.

Let's talk about some tools that you might be able to use. In terms of tech, what we mostly talk about is servers and electricity, because it's an easy one to talk about. You can be in other domains. Obviously, you might be in agriculture, but we can't talk about lots of different ideas. Let's talk about electricity as a real simple one. All of these things are going to change your business landscape. What useful tools are there? Much of the world has talked about all of these other things, I like talking about Wardley mapping. Simon Wardley is a friend of mine. He and I do talk on Twitter a lot. I'm currently working with him on another project. I like Wardley mapping, other people really hate it. This is my take on it.

Basically, Wardley mapping is a way of taking a value chain and being able to visually see that value chain and understand information about it. I'm going to run through his very quick idea of how you can look at a value chain and then assess some of the externalities and how you might see it. Hopefully, you'll be able to see from that, a quick example of how that might be useful. He's got a free online book. You can go and learn it yourself. None of this is proprietary, and you have to hire someone to explain it to you. His idea is a cup of tea. You're trying to sell cups of tea to the public. The public's there. Your business is there. There are the two anchors at the top. You've got your cup of tea. You're buying your cup of tea. Cup, tea, staff, you need your cup, you need your tea to go in the cup. You got stuff to sell to people. You have hot water to go in the cup and the tea. Kettle to boil the water, you need water to go in the hot water, and the power to power the kettle. Power comes at the bottom because it's a long way away from your customers.

Emissions Are Highly Abstracted In a Mature Oil-Based Economy

Actually, that's a really big important thing. In a mature economy, emissions, because that's what happens when you turn the switch on, are abstracted a long way away, huge way away from actual people. In fact, the closest you ever get to actually emitting anything is going in your car. All of your other emissions are pretty much from turning on light switches, or secondary emissions from people delivering stuff to you. Everything else is secondary.

That's why when you talk about climate change, in the environmental side of things, we talk about decoupling, which refers to breaking the link between environmental bads and economic goods. One of the complicated things about that is that our whole economy is essentially based on oil. There are lots of conversations about stopping using oil. If we stop using oil, our economy basically dies. That's the problem with the whole of climate change, which is why we're going to need some really clever technical solutions, and technologies, and things like that over the next 10 to 15 years to go and do something very impressive, which is why we need clever people like you to be thinking about all of these things. Why you need to understand the context before we can talk about some of the other stuff. Basically, power always comes at the bottom in a Wardley map. It's very difficult to get rid of.

If emissions are baked in, basically you can't do anything about it, what can you do? Wardley maps, one of the things you can do is you can start to map finance metrics. You can say, the public give us 55p. The cup costs 5p. The power costs us 5p as well. What happens if you start to map the emissions instead? Ignore the numbers. They're just numbers that don't mean anything at this point. They could mean kilograms of CO2. It doesn't matter what they are. You might have powering the kettle when you create a cup of tea. You might have 8. Then you might have 2 for the cup and 5 for the tea because the tea has been transported to you somehow. You didn't grow it yourself, did you? No, you didn't grow the tea. It's fine. The staff, they have to get there somehow. Then the water gets pumped. Most people don't think about that. You just turn the tap on. Someone has to run a pump somewhere. That has emissions. This is the whole point is that emissions are abstracted. Then the kettle when it heats the water, obviously it's going to use energy. Then you got your business. Adding all of those things up, ends up being that the cup of tea has total emissions of 16.53. That's all we do.

What happens if we change it for something more green? We have renewable power. Then basically a little bit more than 0, but for the purposes of this it's 0. Then we use reusable cups. We're using china. We give nice cups of tea in china. Then we can reuse it in the dishwasher, which means that we're using power for the dishwasher. It's renewable power. There's a little bit of embedded carbon in the making of the China. We use a more efficient kettle. Instead of using that tea, we use a more ethically sourced tea that we ship more slowly using wind power, who knows? I'm making it up as I'm going along, something else. Then the staff, we give them a budget to go on the train or on public transport instead of driving a car. Then we plant a tree, every time someone buys a cup of tea. It takes a bit of carbon out. Of course, that's an offset. We're still emitting carbon, but we're taking a bit of carbon out because we're balancing that one out.

All of a sudden, we have a complete emissions reduction of around -1. We've done a little bit of jiggery-pokery to make it a bit different. You've gone from emissions to fewer emissions, and sorted out. What you've done is you've taken your value chain, and you've actually turned it into something a bit more interesting. You've actually decided. You can see what you can do with your own project. You can understand your own value chain.

Real World - Choosing a Cloud Provider

What about the real world? What about choosing a cloud provider? If all compute was equal, if you were only doing something where you were just using instances? You could do it very easily. You could go, we got three cloud providers that we want to choose from: Google, Azure, and AWS. If sustainability was part of the procedure for choosing a cloud provider, which of those three are we going to choose? I think we might drop one at this point. See what I mean? It's a little map that helps you make some decisions. It's not about getting it perfect. It's a little way of having a visual conversation with the rest of your team to be able to say, at this point, AWS, we'd have to have a very good reason for going with AWS at this point. Maps are useful for identifying and mitigating emissions. They can be used in this way. It's never perfect. Don't try and make them perfect. It's not about that. They're meant to start conversations. It's meant to be used as part of a process of managing your projects.

It's actually really hard to identify emissions in a company. AWS doesn't actually give you emissions. If you go to AWS and say, "Please, can I have my emissions for my account?" They won't give them to you. You can't even offset your own AWS emissions. It's not actually possible unless you are really big and you are basically big enough to go away from them. They might do it at that point, but you'd have to be pretty big. There are companies that are doing stuff like this. If you want to learn a little bit more about emissions, look at scope 1, 2, and 3 emissions. Scope 1 is direct emissions. Scope 2 is indirect emissions. Scope 3 is other indirect emissions. Look it up. It's complicated. Companies like Stripe already do this. They already talk about this. Very large companies that are publicly listed companies already do this. They do scope 1 and 2, at least in ESG metrics, which is why it's useful to understand and to know about these things. If you use Stripe, they're already carbon neutral, which I think is really good. I would use Stripe, for lots of good reasons. This is another one.

What about My Code and Systems?

Let's talk a little bit more about deeper tech. Things that are worth talking about. We've already talked about going green with your cloud provider, talking strategy, and all of those things. Let's talk about my code and system, what can I do? Most climate friendly tech solution is no solution. Don't actually build something that you're going to have to run forever. Please. If you can have something that means you get up and talk to someone, that's actually a little bit easier. Don't actually build tech unless you have to. Why would you? The world does not need more tech all of the time. One thing that keeps coming up in conversation, and has done for years, please, can you turn off all your instances and servers? I spoke to one company recently that worked out that if all of their systems had all of their unused instances, they would be saving 50 MWs of power. By unused, they meant all of the instances that hadn't been accessed for several months in any way. Literally, no network traffic at all. That's a huge amount of power. We're not talking small numbers here. We're talking huge amounts of power. Please, can you just not turn things on and forget about them? That's pretty much what we're talking about here. If you want to build something climate-y and techie that really is interesting. Build some monitoring tools that find out when people haven't turned servers off. Some people will pay you for that because it's worth it. It saves money as well. That's a really good thing.

Stop building climate apps. Please don't build any more climate apps. No one needs them. We have enough ways of figuring out whether or not we're being good and bad for the climate. There are lots of them being built at the moment. Also, code efficiency, I am not interested in whether my code is efficient or not. I don't think you should be. Making something more efficient is not necessarily the right way of going about it. I have a different way of approaching it. It is, "Code like it's 1985." That is, basically, imagine that you're trying to build in 32k. Imagine that your headroom is tiny. Imagine that you're building for devices that are small. Constraints are not bad. It's just that we are so used to building things where we can expand forever. The space is infinite. It doesn't matter if we use 400 libraries for everything. We end up building so that we can expand. If we just accepted the fact that there was a constraint, we would actually build better things. We would use an awful lot less electricity. We would build simpler solutions that were easier to maintain.

Go Serverless

On that point, we should all be going serverless anyway. This is my thing. My background comes from being serverless. I built a startup with serverless. We had 500,000 monthly active users, and a $300 AWS bill. I say that. Half of that was data backup. We spent nothing because nothing was running unless people were using the system. The whole point of serverless is not that you use Function as a Service, because that's not what serverless means. The whole point is you're only using compute when your customers are actually doing something. That's what I think serverless means. That's why it fits with the reduce, reuse, recycle, because you're only using something when you need it. Reducing compute is everything.

I think this is quite important. The sustainability metrics are quite key. Do you know how much energy your project is going to be using over its lifetime? I bet you don't. In fact, I can almost guarantee that you have never even considered it. In fact, the only thing you may have considered is whether or not you're going to be using more servers or less servers over that period of time. What you're going to be doing about scaling. Imagine if you actually thought about, we want to be using as few servers as possible over the lifetime of this project, as an idea from the very beginning of the project. What if you basically said that as a project kickoff? How would you then build something differently because you would? It would be so much better for the environment.

We've already talked a little bit about emissions, scope 1, and 2, and 3. There's limited data in this space. I've been doing a whole bunch of research recently into this space. There is actually limited data on emissions across companies. Many sectors are still trying to catch up in this area. We are really struggling to know where the future of this space will go. We are going to need people who understand data and analytics, and low power data and analytics, and very cost effective data and analytics in this space. I think this is quite a key area for the future.

What about Business Continuity and Disaster Recovery?

What about business continuity, disaster recovery? Slightly more complicated space. Climate change is becoming one of the most important risk factors for the C-level, CxOs all over the place. For a number of reasons, one of them is because their children are going to Climate Strikes. Disaster recovery professionals, not so much, they're beginning to figure it out. Actually, it's not something they're thinking about very much because they don't see it as exceptional. Climate change is changing what exceptional means. It's becoming more normal. It's beginning to change it now. As the O'Reilly scenario shows, it's not about just extreme weather. What are the impacts on family? What are the impacts on employees? Imagine that 20 years down the line, or 10 years down the line, you've got heat waves in places like India, where you may have a massive tech team. What do you do about that? Those heat waves may come sooner? We don't know. Actually, the climate modeling is not perfect. We may get differences. Things may speed up or change in ways that we do not know. You may have to consider this as part of your disaster recovery.

Climate risks are much more than simply sea level rise and a bit of heating, on average across the world. It is a lot more complex than that. How resilient is your organization? Do you actually know what's going on? It isn't just about tech. When we talk about tech we talk about human problems. Most of the time, we're not talking about the complexities of whether or not we know how to do a CI/CD pipeline properly. What we're actually talking about is human problems almost always. Talked a little bit about Wardley mapping, and it's really good for seeing externalities like climate change. It's really good for money as well.

The biggest climate risks that I can see, if you look at BlackRock, which is happening at the moment is capital risk. Most sustainability professionals don't talk tech at all. I've talked to a number of sustainability professionals for a piece of research. Go and be their best friend. I recommend you go and have a one-to-one with your sustainability head at some point in the next month. You will be amazed at how much they love you. Because ignoring the climate crisis isn't an option. This is Mark Carney. He basically said firms ignoring the climate crisis will go bankrupt. He is not wrong. The more I read into this he is not wrong. Climate change is changing our world. There is a new normal coming and it's going to become a new normal again in the not too distant future.


Just because we work in tech does not automatically make us good. Stop imagining that. Change how you think about the world. Change yourself. Go and learn some ethics. Get involved. It's going to take all of us. It's going to take the whole world getting involved. There is a tech group called I recommend go and having a look. I'm on there quite a lot. It's a Slack group, and newsletter, and a whole bunch of other stuff. Otherwise, if you want to find me, this is me. Hopefully, I haven't scared you too much. I think there is an awful lot of hope in this space. I also hope that you've learned a little bit about where this might be going.


See more presentations with transcripts


Recorded at:

Jun 05, 2020