Justin Garrison: Yeah, we’re creating a lot of heat, and we’re consuming a lot of energy. And the closer we can get people to the heat and the energy that they’re having a cause for, I think is important for a lot of people just to understand. [unintelligible 01:12:56.28] carbon footprint calculators, right? We’re like “Oh, how many flights did you have? How far do you drive? [unintelligible 01:13:02.19] How much do you use your phone? How many videos did you scroll on TikTok, and now you’re heating up the Earth because you want to be entertained? Actually, that does have an impact, and it’s crazy when you really think about it that way.

Solana Larsen: Yeah. We called in David Evan Harris, who used to work on the AI responsibility team at Meta, who’s probably a little bit more on the worry side than I am at least, but also has a history in open source, and was able to, I think, also speak in practical terms about what is and isn’t being done at an organization like Meta, and the reasons that he’s concerned, and how it connects to, for instance, social media, which is something that everyday people are really concerned about, right? Like, how is this going to affect, elections and that kind of thing. So that’s something that we looked at.

And then we talked to [unintelligible 00:21:01.14] who is a fascinating researcher, very vocal on Twitter as well, who’s looking for openness in datasets, and who’s been doing a lot of auditing of datasets, and trying to, among a group of researchers who have really tried to engage with companies and try and encourage them to be more responsible about what they do, and do not include in the datasets in regards to hate speech, terrible representations of women, black people, and so forth. And that’s a really important tenant in the way that Mozilla thinks about why it’s necessary for these models to be open. Because the companies time and again have proved that we can’t really entirely trust them. And so open source is like our safety net, in a way.

[21:56] So I think on the spectrum, we’re leaning more towards the open side than the closed side, even taking into account all the things that we know. We talked to Sasha Luccioni from Hugging Face, who does research around climate change and open source models, and just hearing just the simple perspective around – well, simple, but just this idea of if it’s open, if we’re working together in large communities, we might actually have these AI models have a smaller carbon footprint. Because they gobble up lots of energy.

And so they’re the kind of concerns that are outside of this framework of dangerous/not dangerous, where we need to think about what would it mean to have researchers who speak different languages working on these things? What would it mean if we could have a more diverse set of people than the people who work at Open AI working on some of these large language model capabilities?

And then finally, we end up with an interview with [unintelligible 00:22:56.00] from a startup called Nomic, and they’ve made this system called gpt4all, which you may have played with at some point. It’s like an alternative to ChatGPT that you download onto your computer, and you can ask it questions and chat with it offline. And it doesn’t take your data if you ask it not to; it doesn’t take your private information and use it to retrain models. So it’s, again, a different approach to thinking about “Oh, well, what if we did want this stuff to be offline? What if we did want it to be totally compressed, so you could have it on an everyday computer, in a low bandwidth society? What could we do then? How would we be designing differently?” I find that inspiring, but it’s not clear-cut for any of these things. Nobody has the answer. But the important thing is that we need to be really contextual about how we think about these things.

Brett Cannon: You’d be surprised at how complicated forms get when you have two citizenships on these applications and stuff, when you travel abroad… It was a massive hassle, so it’s just simplified my life and just like “I’m good. I’m living here. I’m married to a Canadian. I just don’t need it.” But yeah, it was a lot of fun. But I actually have a Nexus pass, which is really fantastic for getting across down to the US. So getting down’s not a hassle. It’s more time and carbon footprint at this point.

Tim McNamara: [laughs] So timClicks was originally a thing… So I was thinking of like a quirky, pithy handle, and I was clicking around on my keyboard. I was trying everything; it was relatively early in Twitter’s life, so I thought I might be able to get a short handle. timClicks actually was just there because I was clicking on the keyboard, and then eventually, I was like, “Oh, this kind of fits.”

[57:38] The things that that really click for me are an inherent – I’ve kind of got this inherent idea that people are kind of growth-driven, or they are… I’m a real optimist for humanity, and we collectively are facing very significant problems, that are tractable. And this doesn’t mean that I want to be overly prescriptive. I don’t want to be paternalistic with my outlook. I don’t necessarily require that everyone adopts my lifestyle. What I do think is really important is to have a genuine conception of how people are feeling – or really listen, I think is really the thing that I want more people to do. And when someone is disag– so it’s very, very natural in a heated discussion to think “I have thought about this through. I have my position very clear. If you are opposing me, that’s because you are wrong.”

And actually, we’ve known for centuries that two people can look at the same evidence and come to different conclusions. I’m veering off-track here, but what I want to sort of say is that I’ve been really disappointed at how discourse is breaking down, especially I think in the anglosphere. In the English speaking world, we’ve had politicians really corrupt public discourse, and I don’t know how to create something that probably never existed, like a genuine public sphere whereby you could create a space for genuine debate. I think that it’s much more likely that the way forward is through incremental changes, rather than large paradigm shifts to get where we might need to go.

Essentially, given that our society is facing, or like humanity is facing very significant existential threats, to actually reduce the stakes of a lot of the decisions. Like, it’s not like “Oh, we need to go to Mars!” We don’t need these – like, we can survive here at least for another, let’s say, 100,000 years. You think about the technology that we’ve been able to produce in the last 80, to be able to get to spaceflight… Well, you think about projecting that along into, like, let’s say, give ourselves another 100,000 years on Earth, and then we might think around terraforming other planets, or we could colonize other worlds. Because at the moment – or like some other really radical changes that might be required. So “Oh, we need to shift every factory. We need to completely remove carbon.” These all sound like very, very difficult, massive things. But saying, “Oh, actually, tomorrow you can take the bus.” Or “The next meal, or the next time that we cater for an event, we can think about the food that we’re buying, and its carbon footprint.” Or we can think about the ecological sustainability of our purchases; things that we have genuine control for.

I think there’s no single lambda function invocation that’s going to justify Rust. But tens of billions or trillions of invocations - at that sort of scale, things change. And, again, I kind of want to stress that we have problems which are solvable, but they aren’t going to be solved with expecting some government or some massive corporation to kind of make a huge shift. Instead, we can make lots of small changes, and I think that that’s what clicks.

Tejas Kumar: Yeah. Use Deno, save the world. No, but this talks about the reduced carbon footprint of Nx Cloud based on their reduced computation through caching. So they in the last seven days have saved about 400 kilograms of carbon dioxide emissions. 400 kilograms - that’s almost half a ton. In a week. It’s a fascinating read about how they’re doing this. And what I really like about this is it draws attention to responsible computing. Going back to human skills, I think being conscious of our environmental impact is another human skill that’s quite important. And that’s another pillar of impact, if you will, going back to the blog post. We do have an environmental impact at this point that we have to acknowledge. You know why? Because it’s the peak of summer here in Germany, some 100 degrees Fahrenheit, and it just like literally started flooding and bucketing outside, thunderstorms out of literally nowhere. And I’m not used to extreme weather like this, y’all. All of this kind of inspired me sharing this big story of the week.

Gerhard Lazu: [01:10:00.21] Now, that is a problem worth solving, for sure. I hope that this point - we’re not still thinking that it’s not real, I hope. Some will, but that number is less and less by the day. So if we can all agree that it’s a problem worth solving, I’m very curious to hear others how they solve it… Because I know that we have such a big role to play on the infrastructure side, on the apps side, we can make them more efficient, we can autoscale them, or maybe only use as much as we need… We can do so many things there. We can use – maybe be more conscious about the languages that we choose. Some are more energy-efficient than the others.

There was a great talk, and a keynote - the KubeCon EU keynote from Intel, where they ranked the various languages, and they explained the Carbon footprint that they have. Now, I’m not suggesting everyone switch to Rust or C, because they are the most energy-efficient ones… Go is fine, Java too. Surprisingly, Java too; very, very efficient. But there’s something to be said about those, and there’s something to be said about the small choices that we make everyday… Including that burger, which I may not have today, after our talk. I know that’s like the worst food you can possibly have… To maybe cycling more, versus using the car.

Well, thank you for today. It was a great pleasure to tell this story, to re-live a bit of that fun. We did cycle when we were in Valencia, all the way from our hotels, our respective hotels, where we were staying, all the way to the beach. That was a long one. I think it was about like an hour.

Adam Stacoviak: Not to mention the Carbon footprint of shipping those T-shirts to people who don’t even want them, or produced the shirt in the first place.

Forrest Brazeal: You know, it’s changed a lot for me over the years. There was a time when my answer to that would have been very tactical-engineering-focused and saying “Hey, I really just love not having to configure a server” or “I love not having to go out and rack and stack something. I really like just being able to deal with this API and have material provisioned for me.” I think as I’ve progressed, some of the things that I’ve grown to appreciate about cloud are just about the consolidated and centralized nature of expertise that you have access to… And of best practices, and even, honestly, sometimes corporate responsibility.

Something that I really loved - and this was a super-small announcement that came out at Google Cloud Next, where it could have even just flown completely under the radar… This tiny feature where they added a sustainability recommendation into this tool they have called Unattended Project Reminder, where Google’s now able to tell you “Hey, you created this project, but it doesn’t seem like you’ve really been using the resources in it. Do you wanna shut it down to help reduce your carbon footprint?”

A lot of us, for a lot of years, have thought about cloud resource optimization as something that is just about optimizing our build. But when you think about it, it’s actually something that has a societal and planet-scale implication. That gets very interesting.

Brett Cannon: I can only imagine what the carbon footprint of that would have looked like.

John-Daniel Trask: Yes. And I actually touched on it a little bit earlier in our conversation - I struggled a bit with purpose, because I had this goal, which was to own a growing and successful software business. And I achieved that. But a goal is not a purpose, and I’ve confused the two of them. And then when (again) my coach would be like “What’s the purpose?” I’d just keep struggling with this question, because it wasn’t coming to me.

[56:16] And again, I’ll share a Bill Gates one… Their original one was – and it’s funny; they’ve kind of tweaked this, but Microsoft was founded based on Bill Gate’s vision of a computer on every desk and at every home. And the bit that they usually cut from that off the end was “Running Microsoft software.” That was not the bit that was in the public very much. And I was like “Man, you know what - this guy’s amazing. How do you have that clarity of vision and purpose so early on?”

Anyway… Where I’m going with this is that I learned for me it wasn’t to look at what I wanted. It wasn’t to look at it in a positive light. It was actually to figure out something that made me really, really, really angry. Because anger is a signal. Anger itself is almost a useless emotion other than to almost be like an interrupt to your brain, of like “This is important.”

Now, it may be a boundary violation, it might be whatever the heck it is; it’s kind of irrelevant. But it matters. And that was when I was like “Oh, you know what?” This is where I’ve actually focused on the New Zealand piece - I love this country, it’s a great country. I have lived in other countries. I actually lived literally over the road from day one Amazon’s headquarters for three years in Seattle. But that’s where I was like “Why the hell are we wasting all of this Carbon footprint?” New Zealand could have an amazing culture if it actually just hitched its wagon to the transformation that is the software industry around the world? And I’d kind of been a bit tepid about it in the past.

So that was when I started to go “Oh, well if I’m getting (to use the Twitter words these days) triggered by this situation, maybe my purpose is actually to try and resolve that situation, or at least be a contributor to fixing it.” That was the magic for me on kind of going “This is it.” It was figuring out the hill I was willing to die on…

Danya Murali: Yeah, we use a lot of government data. Before I worked at Arcadia, I used to work at the Energy Information Administration, which is the statistical hub of the Department of Energy. So it’s been a cool transition to go from the government agency that collects the data, to being a private entity that uses the data…

But one example I can give about how we use EIA data for good is we try and give our customers an understanding of what their CO2 impact is of their monthly residential electricity use. It’s sort of like being able to keep track of your Carbon footprint, but in a very precise way, that is very much related to your individual usage. We do that using EIA data. So the power of government is you have the ability to go and collect data, and sort of mandate that the data that you’re surveying is getting responded to… Which is great.

One example is EIA has a survey of all the combustion power plants across the U.S. - so all the coal power plants, natural gas, oil… All of those power plants, and how much of each fossil fuel is being used in a year for each of those, and also how much CO2 is being emitted. So using that dataset, and using our individual customer data, we’re able to create individualized forecasts and calculations of how much CO2 is released by using a certain amount of electricity, and then also how much CO2 is averted when you’re able to source your power from places like wind and solar. So there’s a very direct relationship between how we use government data and our data to achieve that.

Marko Saric: Yeah. Basically, when I did the research for this post, I actually went through Google Analytics and I basically counted all the different reports they have, and I counter more than 125 different reports in the left-hand side. All combined, these 125 reports or so, have about 300 different metrics between them. I’m thinking for my own site I probably use 5-10 at most. The other ones - some of them I’ve never heard about or have never looked at them. But still, I’m running the script there that’s kind of calculating these and collecting this data all the time… And I’m using it either never, or maybe once in a while.

I think I’m not the special case here. I think this is quite common. People install Google Analytics because it’s the most popular tool, or it’s something that they’ve told that I need to have… And they have this - this collects 300 different metrics from their visitors, and they use maybe five of them. It’s a waste in terms of the kb load on every website visitor… Thinking in the sense of the climate change even - how to decrease the Carbon footprint of a website, there’s a Carbon footprint calculator. If you decrease something like 50 kb of your site, your score gets so high that you’re something like in the top 10% of the sites that load fast. So it does make a difference, and it’s something people don’t use to look at; and if you don’t use it, then you can remove it. For me personally it makes sense to do that.

Adam Stacoviak: Well, I like your perspective on the push to change around the Carbon footprint, and the flying, and the impact to the environment because of that… Because that was naturally on there. I actually had some opportunities for some of our hosts - we run many podcasts, and a lot of people are involved in Changelog Media things… And so we have lots of places to go and lots of things to do; and a few of them were like “You know what, I’m respecting my desire for a shrunken Carbon footprint this year. I’m gonna resist flying for non-necessary events.” And that’s cool with me, I’m glad you feel that way. And you’re right, it was already kind of on the doorstep of some folks considering that.

Just to dovetail off that a little bit - what’s gonna be really interesting is to see this forced change for everyone, and some sort of analytical look at the climate change because of the lack of all this flying from everyone, vehicles… I drive my vehicle maybe once a week now. Maybe. I’ve got two cars and now I’m thinking like “I don’t even need two cars.” We’ve got two kids, and my wife and I having two cars is just a lot easier for the family to be mobile individually, but now we’re thinking maybe we just need one car.

So there’s a lot less carbon dioxide out there now because of all these things, and just less people doing things, and there might be more data to say “See? If we change in these ways, this is how the environment changes because of that.” Which is great, because that’s data. Data forces change that is true, scientifically true. And if there’s data to support this change, then I think that you might be right, that we might see more virtual events, because we’ll now have more data to make these wiser choices for the environment and the Earth… And our health, too. Smog, and all these things - this is terrible for people’s lungs.

Alexis Richardson: I try to go to as few as possible because of the Carbon footprint impact… But I think that these ones I have to be at at the moment.

Chris Ferdinandi: It’s some sort of an electric vehicle. The other part of this is it’s not just – so I’ve seen blogs that load basically nothing but HTML and CSS, and maybe just a sprinkling of JavaScript for a little bit of interactive pixie dust… I’ve also seen blogs that were built entirely on React. It’s a static site, but the whole front-end is running on React. And if you’re really straining this metaphor that JavaScript is the CO2 of the web, then I think those two different ways of building the same thing have very different carbon footprints. I’ve built entire – the whole front-end is driven by JavaScript before. When people buy my stuff - guides, videos, things like that - they get access to this course portal, where they can access their things, watch the videos, download stuff… The whole thing is JavaScript-rendered. All of the markup is JavaScript-rendered. And I can inline the entire CSS, the base markup and all of the JavaScript to render all of the stuff – just inline it in the HTML file and it gets served in a single 14 kb HTTP request, minified and gzipped.

So even though the whole front-end is built in JavaScript, I’d argue that’s still a much smaller footprint for the end user than if I had to load a framework to do that same kind of rendering. And that doesn’t work for everyone, and it doesn’t work for all sites, and I totally get that… But yeah, there’s kind of like a gradation here; I don’t think there’s any cut and dry “That’s bad. This is good”, and I think you can make a clear argument that for certain large-scale sites using a framework does have some legitimate benefits. I think we lean on the idea that “Oh, at scale, at scale” way too often. I think a lot of the things people throw that argument at - they’re not really building at the kind of scale that these tools were designed for. But that’s a different conversation.

Nicholas Harris: Yeah, so training is R&D mode. You’re building a tool that takes inputs, and then tells you something about those inputs. If that’s all we ever did in machine learning, no companies would fund this stuff, because the ultimate goal of building those models is that you want to use them for something. If you’re Google or Facebook, these neural networks get deployed at massive scale, and their job is to take a user query, run a neural network on the query, give you results. And sometimes there are lots of neural networks in that chain, and that is where the overwhelming majority of the energy footprint of AI will be over time.

Training is a really big deal. You gave the example - I think MIT Tech Review did that analysis that you were quoting, where it was like five cars over their whole lifetime is how much energy the training was using, and I think the equivalence was in carbon emissions. So yeah, inference is deployment, which is where most of the scale is, and training is just doing the R&D, that lets you do that thing that makes you money. Because the training stuff doesn’t make you money.