ngould

7 ms·

31.
by ngould 10y ago
My understanding is the CUDA interface is for NVIDIA GPU's only. There are non-CUDA API's for other types of GPU's, but they don't necessarily work out of the box with neural network libraries like Tensorflow yet.
32.
by ngould 10y ago
Cool!
33.
by ngould 10y ago
Mortgage backed securities were not risky because they were diversified. They were risky because nobody fully understood how undiversified they were, especially at the lower grade tranches. Meanwhile, LC was alive in '08, and you can l
34.
by ngould 10y ago
Historically people have made money on Lending Club, and lost it in Vegas. I'll stick with Lending Club.
35.
by ngould 10y ago
The star schema is still very useful, even though it wasn't invented in the last six months.
36.
by ngould 10y ago
If you're interested in defining dimensions using tabular data in JS, check out crossfilter.js. https://github.com/square/crossfilter
37.
by ngould 10y ago
I'm surprised you didn't mention the mortgage interest tax deduction as a contributor to sprawl. I.e. federally subsidized home ownership. Talk about reasons to ditch your downtown rental and borrow against a grassy slice of subur
38.
by ngould 11y ago
Very cool, thanks for the update.
39.
by ngould 11y ago
Hey pkrein, I am a Segment fan and startup-scale user of your warehouse service. The warehouse service is super convenient, but its main shortfall for us is that you don't offer historical backfilling of data more than 60 days before t
40.
by ngould 11y ago
Word on the street is Periscope costs $1000/mo for unlimited users, up to 1B rows.
41.
Activate: Tech and Media Outlook 2016(slideshare.net)
2 points by ngould 11y ago|0 comments
42.
by ngould 11y ago
Cool stuff. Curious to know what systems these containerized tasks are pulling data from. Does adroll consider let those containers access production database instances? Or are they backed by non-production systems? (EDIT: I understand that