OKAY SO TIME TO WHIP OUT THE DATA SCIENCE TOOLS
We can definitely do better than copying text 
a quick look at the ink counter's source reveals that the ink number is in plain text for easy scraping
booooo Tapas blocks python access sources so I can't use requests... so I had to refresh on bash
So now I just watch -15 inkbash and ping every 15 seconds sweet
... At this point I spent about 3 hours trying to develop a cool machine learning regressor so I could look smart and cool but tensorflow absolutely hates me and I have no idea how to properly work 1.12 lmao I use 1.3 for work and it's terrible. I got mean square error rates of 25000. I figured out normalization could get the optimizer to stop freaking out but then I'd lose my precision. Eh... reinforcement learning is more my thing anyway...
So scipy linregress it is. Here's the code
import numpy as np
import matplotlib.pyplot as plt
from scipy import stats
data = np.empty([0]) #lol don't judge me
label = np.empty([0])
with open("/home/sol/bashout") as f:
for l in f:
line = l.strip()
t = line[:10]
i = line[-14:-4]
i2 = i.replace(",", "")
data = np.concatenate((data, [(float(t))]), axis=0)
label = np.concatenate((label, [(float(i2))]), axis=0)
print(len(data))
slope, intercept, r_value, p_value, std_err = stats.linregress(data, label)
plt.plot(data,label, 'o', label='orig')
plt.plot(data, intercept+slope*data, 'r', label='fitted')
plt.legend()
plt.show()
print(intercept+slope*1542870000) # that's 12am PST in epoch time
If I use all the data I collected since (625 pings == 2.6 hours or 156 minutes) I get
Expected prediction is 16,715,755 ink
Ok, but honestly, it looks really bumpy and the right end is starting to taper off (makes sense, people are less enthusiastic to donate since we met the 15 million goal, and it's getting late)
So I re-calibrated to the last 125 points (about 31 minutes, so between 10:20-10:50 est)
So it really is tapering off. I could probably get even more accurate if I also waited until 12:30 am EST (10:30 pst) but I'm tiiiired 
AND SO: my final prediction as of 11:00 EST is --drumrolls--
16,375,435
Happy thanksgiving!! 
