As much as I appreciate @awscloud's "spend all of your money to find out why you're spending all of your money" approach here, this is approaching "step 17, overpower the wolf" levels of hilarious complexity for disappointing outcomes.
This blog post has, and I am being completely serious, 56 steps you must undertake to arrive at the disappointing outcome.
At that point your bill is high, your staff time costs is way higher, your predictions are no more accurate than they were before, and this is your rebuttal to the reInvent Machine Learning keynote:
• • •
Missing some Tweet in this thread? You can try to
force a refresh
