This is something that I couldn’t get out of mi mind, when I tested the new version of my Twitter AI robot (v4). I witnessed 800 new followers in the first 2 hours in a brand new Twitter account.
I achieve this high number by applying parallel computing, with the current level of hardware available now, it is possible to exploit the multi threaded processors available now in common computers, yes in a single machine.
But as the number of followers approached the 10,000 level, the number of followers per hour kept on going down to under 200, that is a loss of efficiency or more than 50%! This is a little too much for me to bare.
So in the next day or so I was thinking… what happens if I apply the same idea (parallel computing) to the other end, to the servicing side. What happens if I service multiple accounts?
This is what I am testing now and these are the preliminary results:
| Started | Age | Account | Following | Followers | Average | Seconds | Time | Date |
| 06/24/14 | 33 | hits1k | 15 | 23,593 | 707 | No longer in the pool of service | ||
| 07/22/14 | 5 | hits1000 | 4 | 10,443 | 1,942 | 195 | 08:45 | 07/27/14 |
| 07/26/14 | 1 | hits1k_ | 242 | 1,801 | 1,308 | 197 | 08:41 | 07/27/14 |
| 07/26/14 | 1 | hits1k_1 | 211 | 1,696 | 1,232 | 200 | 08:38 | 07/27/14 |
| Totals | 33 | 472 | 37,533 | 1,125 | 59.09% | Improvement | ||
Well, the idea seems to be working, what do you think?
Accounts:
The first account (hits1k), was removed from the service when it reached 26,000 followers to see what happens with the acquired followers, a reduction was expected, the first 8 hours were scary, over 1,000 left, 5 days later seems to be ok.