"Top 10 data mining algorithms in plain English" lives up to its name http://t.co/pWqxMOorZO (via @dataelixir)
(Unedited) recording from my Halogen talk today #purescript https://t.co/6Je1CouoAv
New (epic) blog post on "The Unreasonable Effectiveness of Recurrent Neural Networks" http://t.co/OKkaxvIGnS was immense fun to write
If training vanilla neural nets is optimization over functions, training recurrent nets is optimization over programs.
The basic idea is that there's a lot of wisdom in these essays, but unfortunately Paul Graham is a relatively slow generator. Wouldn't it be great if we could sample startup wisdom on demand?
In case you were wondering, the yahoo url above doesn't actually exist, the model just hallucinated it.
.@Google introduces inexpensive virtual machines as low as $0.01 per CPU core hour