I can't believe it took this long for someone to make a Bitcoin-mining space heater

Originally published at: I can't believe it took this long for someone to make a Bitcoin-mining space heater | Boing Boing


Given the drop in crypto prices, I suspect it won’t remotely pay for itself anymore and it’s $20 a month, period. You’ll just be warming your house in the most inefficient way possible - but presumably the seller makes money off it up front and by skimming off the crypto transactions, so they still have an incentive to sell the things.


It’s going to be safer than a space heater as the GPU typically doesn’t get hotter than 100C (200F) whereas a space heater has an exposed resistive wire coil inside. Don’t see how you’ll recover your investment in the months when you don’t need heating. The energy you put in isn’t being offset by doing anything else useful and you still have to cool your home. You may as well play video games all day. Hammering an expensive GPU 24/7 means you’ll probably have failures before you recoup the cost.


I mean, you shouldn’t be heating your home with a space heater…but if you do need one it shouldn’t be a problem, right? Inefficiency is how you make heat, so it’s pretty much the only case for mining crypto that doesn’t seem completely horrible to me. Although you could still always be doing useful math instead, like protein folding or phylogenetic trees.


You’re right, it didn’t take this long to come up with the idea. It’s been done since at least 2017, with some entrepreneurs trying to make a business of it in 2019, in Siberia where it makes a lot of sense. Obviously Siberia has plenty of cold weather and I assume they subsidize the electricity there.


Yeah this also. Tons of distributed math problems could be attacked by this idea. Everything from protein folding, as you mentioned, to weather prediction, fluid modeling, fusion plasma modeling, training neural networks, all kinds of tasks.


This kind of shit is why I still can’t upgrade my damn graphics card.


Yes, relatedly, this is why California making incandescent bulbs difficult to source (hard to buy them even online here) - while most certainly of all the finest intentions, really does miss the mark both from an energy and carbon perspective - in many cases (esp areas served with hydro generation). Options for residential heating are gas (also being phased out) or electric - and watts are watts, so a 100W tungsten bulb is a nice little boost heater in many areas during many parts of the year.


I thought that GPU mining of Bitcoin had been a losing proposition even before the recent crypto crash, simply because custom ASIC chips have mined bitcoin far faster and more efficiently than GPUs for years now?


The last few years the only serious mining stories I’d heard in the West involved incredibly sketchy stories of stealing electricity. In effect, with the ‘difficulty’ increasing with more miners, you’ve got miners bidding on the cost of mining a block.

When a group of miners is paying nothing, yeah, that’s going to push the price below what someone actually paying the bills can afford.

The sketchy stories included:

  • normal illegal splices bypassing the meter.
  • terrifyingly underrated cables running to the shed in ‘bills included’ rentals.
  • malware.
  • malware, at work!

Also, I believe China’s had cheap hydro since one of the dams came online so $/MWh is less over there. It’s all internet, so whereever the cheapest mining is, mining will go. Unfortunately Iceland paused looking at exporting stored energy because it was economical to use the surplus geothermal energy to mine bitcoins. Some of that might pivot into neural net servers, but more likely that hardware is fried.


If they could design it to apply the heat to an absorbtion refrigerator it could be used year around.

The heating element in a typical RV fridge is 250 - 300 watts but it needs around 300+ degrees F to work well.

A smaller absorbtion cooler would probably work well.


An LED bulb just won’t work in an old school Easy-Bake Oven.


'zactly! - also the very odd 20V MR16 halogens that go with the variac fiber light sources in the scopes in our lab - (making the case to smuggle those over state lines to the purchasing dept. was fun)


I also assumed this, GPUs are no longer useful for Bitcoin. Maybe they are for some other cryptos? For BTC, it’s all about the hashes per joule, and GPUs aren’t designed for that (they are designed for floating point matrix ops) and can’t compete with ASICs.

Anyway, this all would make sense only in some cases where electricity is cheap and people need to heat their houses a lot and crypto prices are favorable relative to either ASIC or GPU prices (whatever is suitable for mining).


AND where electricity is the main choice for heating.

1 Like

Landsvirkjun has been throttling power to crypto miners and aluminium plants and refusing any new applications for crypto because of a hydro power shortage in SW Iceland. The mining industry has been using more power than the entire nation’s domestic consumers, and it has been paying subsidised rates, unlike the citizens. Oh and the big companies don’t pay any measurable amount of tax in Iceland.

Power is cheap in Iceland - but it’s stupidly cheap for industry and the government keeps subsidising new energy generation projects that rip up the landscape to satisfy their needs.


BB Begs to differ:

1 Like

All electronics are technically a space heater? Maybe not a good one but definitely one.


buying a new computer is not the most efficient way to source a heater though

( is what i assume he meant )

in a degenerate sense, of course, all electric heaters are equally efficient—a watt is a watt


I’m guessing the efficiency is crap even for a space heater. And given the (low) value of crypto now, the difference can’t be justified by the $20 return (much less the enormous up-front cost).


All resistive-type electric heaters, yes. But electric-powered heat pumps can provide more than one watt of heating to a space for each watt used to power it. Under the most ideal conditions a heat pump can transfer about 300% more heat energy than the electric energy that it consumes. Most of the time it’s less than that but still almost always greater than an old fashioned resistive heater.