exp1orer

I tried switching from Spotify to Youtube Music but the "download for offline" functionality is incredibly broken -- often doesn't recognize that it's on wifi to start downloading, multiple minutes to download an album that Spotify downloads in a few seconds, and for a longer playlist the pause/resume when you go off wifi doesn't work at all.

It might be too soon to call it abandoned, but I was very intrigued by the Austral [1] language. The spec [2] is worth reading, it has an unusual clarity of thought and originality, and I was hoping that it would find some traction. Unfortunately it seems that the author is no longer actively working on it.

[1] https://austral-lang.org/ [2] https://austral-lang.org/spec/spec.html

if anything is fungible, surely mushrooms are...

This is really cool!

If the author is around, I notice in the README you mention the GNU units program, which I use quite a bit. I'm curious if you've made any notable divergences from it?

a few clarifications (in the industry but not involved with this deal):

* carbon removal credits are a subset of carbon credits

* they are generally considered higher-quality than most other credits (which are "avoided emissions"). This is because, for example, turning on a direct air capture machine, is clearly something that would not happen without the sale of carbon credits.

* there's not always a clear line between carbon removal credits and non-removal (ie "avoided emissions").

* unfortunately the carbon credits that have come under the most fire (nature-based solutions like forestry) are also, technically, closer to being "carbon removal" -- and some sellers play up that ambiguity, to their advantage.

I love this essay. Anna Weiner's book Uncanny Valley has a great line about it as well:

> "it's like no one even read 'The Tyranny of Structurelessness,'" said an engineer who had recently read The Tyranny of Structurelessness.

I agree that this is likely missing data systematically.

However note the difference between carbon offsets vs carbon removals. I don't know what kinds of offsets Google bought but given that they started buying them >15 years ago, they were probably not removals.

I actually hesitated before posting for exactly this reason, and I don't think you're wrong to be sensitive to this. But you'll notice that there's actually no real political content in the link, nor was there any political discussion in the thread, so I think your response is a little bit of an overreaction.

Ha, no, just a fun exercise sparked by a conversation with a friend. It also doesn't take the road network into account, just county adjacencies.

Vaguely related, a few years ago I made a map of "how to drive from SF to NYC while passing through the minimum number of Republican-voting counties": https://github.com/louispotok/blue-road-trip

(OP but not author: the original seems to only have releases for windows, there's a python fork here: https://github.com/dlenski/wp2git )

This was the bit that got me:

> People have this aspirational idea of building a vast, oppressively colossal, deeply interlinked knowledge graph to the point that it almost mirrors every discrete concept and memory in their brain. And I get the appeal of maximalism. But they’re counting on the wrong side of the ledger. Every node in your knowledge graph is a debt. Every link doubly so.

If I'm reading your comment correctly, you may not be aware that methane abatement is also eligible for carbon crediting programs, which operate under the framework of "global warming potential" (GWP) to translate different gases into "CO2-equivalent tons" (tCO2e).

Maybe virtual assistant?

1. Incredible line.

2. That 6% is based on the most recent IPCC report which estimates 1.4 GT from HFCs, and another ~1.4 from CFCs+HCFCs, so 2.8 total. OWID seems to get its data from climatetrace, I haven't dug into their data, but if I'm reading that correctly it looks like that is only energy usage from manufacturing, rather than direct emissions of the gases. Looks like maybe this breakdown just ignores refrigerants, which is unfortunately common.

Tradewater is great, they are one of the original inspirations for this approach. The key differences are geography (they are mostly US-focused) and that they pursue stockpiles of gas where we go after the emissions from end-of-life equipment. So their counterfactual is slow leakage over time, and our counterfactual is immediate venting.

It's 2000x worse on a per-pound basis, using the Global Warming Potential (GWP), and 6% of total emissions on a CO2e basis. They are separate facts which are mentioned to serve different purposes: 6% is to emphasize the scale of the overall problem, and 2000x is to understand the leverage of this approach (small amount of material to handle for the impact).

Yeah, R600a is a so-called "natural refrigerant" with a GWP that rounds to zero. Not even close to in scope for us -- the lowest GWP gas we touch is R32, with GWP 750.

Your bio suggests you know more about the chemical details than I do, but we do have

1. A trial burn done by an independent lab to make sure that under normal operating conditions, when destroying refrigerant the levels of those + other chemicals are below certain thresholds

2. automatic monitoring and shutoff mechanisms if the kiln deviates from normal conditions

That is a super interesting idea! Hadn't considered it, have you seen anyone doing this with carbon credits before?

Good questions, we spend a lot of time thinking about how to ensure the actual additionality of what we're doing. The most important consideration - we pay per kg, and less than the market rate of new gas. So a technician that was recovering gas before to reuse it would be taking a loss if they sold to us and then went out to buy new gas.

Good idea! I looked into this when I first started thinking about the problem. MethaneSAT is launching this year, so I thought maybe we could do the same for refrigerants. Unfortunately I don't think it would work:

1. Methane is 1900 ppb in the atmosphere[1], refrigerants overall are ~12.5 ppb -- so a lot less out there to detect. 2. I think methane leaks tend to be from single, continuing, large point sources -- coal mines, oil wells -- whereas refrigerants are usually from these distributed small point sources which happen once. So by the time you detect it it's too late to intervene.

[1] https://www.noaa.gov/news-release/increase-in-atmospheric-me...

We destroy it in an active cement kiln! The Montreal Protocol has a Technology and Economic Assessment Panel outlining approved destruction options[1] and this is one of their approved technologies. It's great for a few reasons:

1. Already operates at negative pressure with a high enough temperature and long enough residence time 2. Alkaline environment neutralizes the HF and HCl that are produced when the refrigerant burns. 3. Already consuming massive amounts of energy so the marginal energy use is negligible. 4. Allows use to use an existing facility instead of building our own -- great for developing countries where building infra is harder (but cement plants are everywhere).

[1] https://ozone.unep.org/sites/default/files/2019-04/TEAP-DecX...

Good question!

In theory, reclaiming gases that are still unrestricted for production/import is at least as good as destruction from a climate perspective. However, virgin refrigerant is really cheap until import bans take hold -- so there is never a point where it is both economically worthwhile and impactful for climate. In theory you could use credits to boost the economics around reclaim but you end up with a very messy additionality story. My sense is that most reclaimers are very low-margin or even loss leaders for the companies that produce/sell the gas!

Anyway the short answer is that it's harder for us to figure out in this first push, but we do intend to look into it more closely as we expand.

Precisely because credits are not really a commodity, so some buyers have preferences about what credits to buy and how much they are willing to pay. There are some subset of buyers and sellers who pretend that all credits are the same, but usually that's an excuse to pay for the cheapest possible credits (which in many cases achieve nothing).

Yeah that's definitely an issue! It's probably a bit less tractable than what we're dealing with because

1. Scrappers likely don't have the skills, equipment or inclination that AC technicians do.

2. Those self-contained units typically have a lot less refrigerant than the split AC systems (or large chillers) we see, which require venting in order to remove once installed. A refrigerator might have only 30 grams of charge, vs a split system with 2-3 kg.

It varies by country but I'm not aware of any country that has fully phased out HFCs -- the EU I believe is already imposing import quotas but it is still legal to use.

There is a link to the IPCC report a few paragraphs down where you can read about this and other statistics to your heart's content.

There is definitely a lot of bullshit out there but when companies decide to pay for climate mitigation, even as a marketing ploy, I think that is net good.

Many companies are certainly looking to buy the cheapest credits they can find but there are promising indications that things are changing, led by companies like Stripe and Shopify.

We are 10x cheaper than the high-quality carbon removal that for example Stripe Climate is purchasing. But we are 10x more expensive than the "low quality" credits that I describe in the post. So overall there is a 100x differential between what is allegedly the same ton of carbon, based on perceived quality and other factors.