4 ms·

I highly doubt this. He cites 40MB of downloads plus high CPU usage. Mining bitcoin won't need 40MB of download. Plus it has an ROI of 0. Even with free adspace

I highly doubt this. He cites 40MB of downloads plus high CPU usage. Mining bitcoin won't need 40MB of download. Plus it has an ROI of 0. Even with free adspace this would not be worthwhile. More likely is someone just fucked up their ad.

kardos 9y ago
I only somewhat doubt this. Someone was thinking "1e-8 bitcoins per pageload, times 100000 people per day, ...". If they sort out how to leverage the GPU via javascript then they might make a little profit.
- delinka 9y ago
  WebGL shaders, right?
- mikeash 9y ago
  I don't think it's even that high. It looks like most GPUs are well under 1GH/s. If we assume 1GH/s, then at the current difficulty and exchange rate, if a page load gives you one minute of computation on average (wild guess, but seems reasonable), you're looking at about $5e-7 dollars per pageload. That's 5 cents/day if you get 100,000 people per day. But this isn't exactly an argument that it can't be happening. They might have failed to understand just how futile it would be.
wyldfire 9y ago
The ROI is likely nonzero for some coins out there. Bitcoin's difficulty would be too high to make much money even if the ad were served to a very large number of computers. But other non-SHA coins (CryptoNight-based PoW, e.g.) or low-difficulty coins might be worthwhile. > Mining bitcoin won't need 40MB of download. Are you sure? I would think something like cgminer compiled to wasm/asm.js could be a 40MB payload. For more context, OP states: > "checked sources and discovered it was a Bitcoin add in iframe, looked at JS it used" > "It was also an ad for a btc exchange/wallet thing, so maybe the mining was just to recoup some cost?" The nebulous "sources" referred to might be something credible. I imagine bitcointalk.org may have a thread on malicious ads like this.
- SparkyMcUnicorn 9y ago
  Also, he says "It may have been another currency, I can't be certain. Should have generalised" https://twitter.com/adamjogrady/status/902533084767072258 https://twitter.com/adamjogrady/status/902533084767072258
- LyndsySimon 9y ago
  > The ROI is likely nonzero for some coins out there. If it's more profitable to mine $ARBITRARY_ALT_COIN than Bitcoin, then people would mine $ARBITRARY_ALT_COIN until the profitability equalized. > Are you sure? I would think something like cgminer compiled to wasm/asm.js could be a 40MB payload. There's no need for that: https://github.com/jwhitehorn/jsMiner https://github.com/jwhitehorn/jsMiner
  - wyldfire 9y ago
    > If it's more profitable to mine $ARBITRARY_ALT_COIN than Bitcoin, then people would mine $ARBITRARY_ALT_COIN until the profitability equalized. That's a reasonable theory but I don't think it holds. Given that (1) $ARBITRARY_ALT_COIN aren't typically well-known, (2) setting up mining may require different hardware/software to enable computing this PoW, that pool, etc, (3) coins and pools have interesting reward & difficulty adjustment algorithms, some better designed to deter coin-hopping. On some coarse-enough scale the general claim of equalization is reasonable -- the differences will be minimized over time. But I have seen times when you could make 5-10% more by mining $ARBITRARY_ALT_COIN (and perhaps taking the altcoin to exchange for bitcoin). > There's no need for that: https://github.com/jwhitehorn/jsMiner https://github.com/jwhitehorn/jsMiner shrug, seems like we're on a tangent. This thread was speculating about whether the claim of an ad-served-miner exists. The fact that a small JS miner exists for one PoW doesn't seem to rule out the original claim, nor does it rule out the possibility that a wasm/asm.js cgminer might exist.
    - LyndsySimon 9y ago
      > But I have seen times when you could make 5-10% more by mining $ARBITRARY_ALT_COIN (and perhaps taking the altcoin to exchange for bitcoin). Right - maybe you could even make 100% more with an alt, for a limited period. My point is that you'd need to be able to make 20000x as much as you could make with Bitcoin before you reach parity with using traditional ads. > shrug, seems like we're on a tangent. This thread was speculating about whether the claim of an ad-served-miner exists. The fact that a small JS miner exists for one PoW doesn't seem to rule out the original claim, nor does it rule out the possibility that a wasm/asm.js cgminer might exist. Gotchya. I was just trying to show that giving access to computing resources instead of viewing ads is not currently anywhere close viable.
Rjevski 9y ago
Well given that mining bitcoin in a browser isn't even remotely profitable, someone crazy enough to do it anyway would also be crazy enough to not give a fuck about their code's efficiency and download 40MB's worth of data.
- Retric 9y ago
  Mining on GPU's is generally not profitable, but if you don't pay for the GPU or electricity then numbers look different. Especially if you get some people that keep active tabs open for days. Really the ROI is vs a 40MB download and a few days of R&D.
  - Rjevski 9y ago
    I doubt browser-based mining is even using the GPU at all.
    - Retric 9y ago
      HTML5 gives access to GPU resources. EX: https://developer.mozilla.org/en-US/docs/Web/API/window/requestAnimationFrame https://developer.mozilla.org/en-US/docs/Web/API/window/requ...
- rtkwe 9y ago
  Depends on how much each ad cost them. The main reason that ASICs are the only profitable way to mine is because of electricity costs and by running this ad they're shunting that cost onto the user of the site. In place of that is the cost to serve an ad which can be really cheap. I could see the math working out buying a lot of really long tail terms super cheap.
  - the_stc 9y ago
    CPU mining is not profitable even for free. Assume they get 5 minutes of computation per ad at 100H/s. They need billions of ads a day just to get 1GH/s, which is around 48 cents a year.
hdra 9y ago
They don't need any ROI from the mining though, they're paying for the ads to their exchange. They only need the ROI on the visitors to their website. The mining is just an added bonus.
ringaroundthetx 9y ago
It has an infinite ROI if you are using other people's processing and electricity.
- paulpauper 9y ago
  if the code hurts performance to such a degree that it is uncovered and uninstalled, then no
  - ringaroundthetx 9y ago
    To say and defend that you would have to either not know how ROI works, or how mining pools work, or both.... I guess I'll help: The costs in this case would be the time and cost of programming and bundling the ad, as well as the ad spend of getting the ad shown. The earnings would be in cryptocurrency. Any non-zero amount of cryptocurrency earned would mean you have a return .... on investment, and any cryptocurrency amount now or in the future that is valued at more than your costs means you have an ROI greater than zero. Mining pools count shares of work toward the pool itself finding a block, based on processing power contributed to a pool. So it doesn't matter if users uncovered and uninstalled your mining-ware after 5 minutes in the most technical savvy outlier case, the shares of their processing power was already submitted to the pool and that account got paid. So wouldn't it be hard or almost impossible to say they have an ROI of zero? They are probably making a good amount of money.

kardos 9y ago

I only somewhat doubt this. Someone was thinking "1e-8 bitcoins per pageload, times 100000 people per day, ...". If they sort out how to leverage the GPU via javascript then they might make a little profit.

delinka 9y ago
WebGL shaders, right?
mikeash 9y ago
I don't think it's even that high. It looks like most GPUs are well under 1GH/s. If we assume 1GH/s, then at the current difficulty and exchange rate, if a page load gives you one minute of computation on average (wild guess, but seems reasonable), you're looking at about $5e-7 dollars per pageload. That's 5 cents/day if you get 100,000 people per day. But this isn't exactly an argument that it can't be happening. They might have failed to understand just how futile it would be.

wyldfire 9y ago

The ROI is likely nonzero for some coins out there. Bitcoin's difficulty would be too high to make much money even if the ad were served to a very large number of computers. But other non-SHA coins (CryptoNight-based PoW, e.g.) or low-difficulty coins might be worthwhile. > Mining bitcoin won't need 40MB of download. Are you sure? I would think something like cgminer compiled to wasm/asm.js could be a 40MB payload. For more context, OP states: > "checked sources and discovered it was a Bitcoin add in iframe, looked at JS it used" > "It was also an ad for a btc exchange/wallet thing, so maybe the mining was just to recoup some cost?" The nebulous "sources" referred to might be something credible. I imagine bitcointalk.org may have a thread on malicious ads like this.

SparkyMcUnicorn 9y ago
Also, he says "It may have been another currency, I can't be certain. Should have generalised" https://twitter.com/adamjogrady/status/902533084767072258 https://twitter.com/adamjogrady/status/902533084767072258
LyndsySimon 9y ago
> The ROI is likely nonzero for some coins out there. If it's more profitable to mine $ARBITRARY_ALT_COIN than Bitcoin, then people would mine $ARBITRARY_ALT_COIN until the profitability equalized. > Are you sure? I would think something like cgminer compiled to wasm/asm.js could be a 40MB payload. There's no need for that: https://github.com/jwhitehorn/jsMiner https://github.com/jwhitehorn/jsMiner
- wyldfire 9y ago
  > If it's more profitable to mine $ARBITRARY_ALT_COIN than Bitcoin, then people would mine $ARBITRARY_ALT_COIN until the profitability equalized. That's a reasonable theory but I don't think it holds. Given that (1) $ARBITRARY_ALT_COIN aren't typically well-known, (2) setting up mining may require different hardware/software to enable computing this PoW, that pool, etc, (3) coins and pools have interesting reward & difficulty adjustment algorithms, some better designed to deter coin-hopping. On some coarse-enough scale the general claim of equalization is reasonable -- the differences will be minimized over time. But I have seen times when you could make 5-10% more by mining $ARBITRARY_ALT_COIN (and perhaps taking the altcoin to exchange for bitcoin). > There's no need for that: https://github.com/jwhitehorn/jsMiner https://github.com/jwhitehorn/jsMiner shrug, seems like we're on a tangent. This thread was speculating about whether the claim of an ad-served-miner exists. The fact that a small JS miner exists for one PoW doesn't seem to rule out the original claim, nor does it rule out the possibility that a wasm/asm.js cgminer might exist.
  - LyndsySimon 9y ago
    > But I have seen times when you could make 5-10% more by mining $ARBITRARY_ALT_COIN (and perhaps taking the altcoin to exchange for bitcoin). Right - maybe you could even make 100% more with an alt, for a limited period. My point is that you'd need to be able to make 20000x as much as you could make with Bitcoin before you reach parity with using traditional ads. > shrug, seems like we're on a tangent. This thread was speculating about whether the claim of an ad-served-miner exists. The fact that a small JS miner exists for one PoW doesn't seem to rule out the original claim, nor does it rule out the possibility that a wasm/asm.js cgminer might exist. Gotchya. I was just trying to show that giving access to computing resources instead of viewing ads is not currently anywhere close viable.

Rjevski 9y ago

Well given that mining bitcoin in a browser isn't even remotely profitable, someone crazy enough to do it anyway would also be crazy enough to not give a fuck about their code's efficiency and download 40MB's worth of data.

Retric 9y ago
Mining on GPU's is generally not profitable, but if you don't pay for the GPU or electricity then numbers look different. Especially if you get some people that keep active tabs open for days. Really the ROI is vs a 40MB download and a few days of R&D.
- Rjevski 9y ago
  I doubt browser-based mining is even using the GPU at all.
  - Retric 9y ago
    HTML5 gives access to GPU resources. EX: https://developer.mozilla.org/en-US/docs/Web/API/window/requestAnimationFrame https://developer.mozilla.org/en-US/docs/Web/API/window/requ...
rtkwe 9y ago
Depends on how much each ad cost them. The main reason that ASICs are the only profitable way to mine is because of electricity costs and by running this ad they're shunting that cost onto the user of the site. In place of that is the cost to serve an ad which can be really cheap. I could see the math working out buying a lot of really long tail terms super cheap.
- the_stc 9y ago
  CPU mining is not profitable even for free. Assume they get 5 minutes of computation per ad at 100H/s. They need billions of ads a day just to get 1GH/s, which is around 48 cents a year.

hdra 9y ago

They don't need any ROI from the mining though, they're paying for the ads to their exchange. They only need the ROI on the visitors to their website. The mining is just an added bonus.

ringaroundthetx 9y ago

It has an infinite ROI if you are using other people's processing and electricity.

paulpauper 9y ago
if the code hurts performance to such a degree that it is uncovered and uninstalled, then no
- ringaroundthetx 9y ago
  To say and defend that you would have to either not know how ROI works, or how mining pools work, or both.... I guess I'll help: The costs in this case would be the time and cost of programming and bundling the ad, as well as the ad spend of getting the ad shown. The earnings would be in cryptocurrency. Any non-zero amount of cryptocurrency earned would mean you have a return .... on investment, and any cryptocurrency amount now or in the future that is valued at more than your costs means you have an ROI greater than zero. Mining pools count shares of work toward the pool itself finding a block, based on processing power contributed to a pool. So it doesn't matter if users uncovered and uninstalled your mining-ware after 5 minutes in the most technical savvy outlier case, the shares of their processing power was already submitted to the pool and that account got paid. So wouldn't it be hard or almost impossible to say they have an ROI of zero? They are probably making a good amount of money.