Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
mhluongo
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
60 ms
·
691.
▲
by
mhluongo
13y ago
Maybe the business they're validating is an HN-specific PR firm :)
692.
▲
by
mhluongo
13y ago
There are plenty of super-early or even idea stage companies that apply, right? Seems relevant in that case.
693.
▲
by
mhluongo
13y ago
Interested, and also in Atlanta. Email is handler @ gmail.
694.
▲
by
mhluongo
13y ago
Thanks. In touch with a couple lawyers now.
695.
▲
by
mhluongo
13y ago
Retail.
696.
▲
by
mhluongo
13y ago
I'm definitely just wanting background- I don't think I'm a lawyer.
697.
▲
by
mhluongo
13y ago
I don't know, these are pretty good answers. My concerns: the law, fraud, and finding a balance between this and my full-time job, in that order. Overarching theme- how to make this work while minimizing my personal risk.
698.
▲
by
mhluongo
13y ago
Ha, thanks, I'm looking forward to the docs. Re: fraud I've definitely been disillusioned that I can handle the tide myself, but OTOH I imagine fraud prevention is a lot like security- make sure the wall is higher than the water.
699.
▲
by
mhluongo
13y ago
I'm not accepting registered cards, which changes the fraud vectors quite a bit. No auto-reload, etc.
700.
▲
by
mhluongo
13y ago
Hand-verification right now before payout, though I'm locking it down further now.
701.
▲
by
mhluongo
13y ago
One of the steps to de-risk the cards will be a waiting period- new cards won't be accepted unless they've been activated X days/months. That should seriously lower the chance of stolen cards getting through.
702.
▲
by
mhluongo
13y ago
That was my thought too, but this is a common reaction so I'm definitely seeking clarification.
703.
▲
by
mhluongo
13y ago
I've been hand-verifying the transactions so far to try to keep abreast of fraud. I've also taken some automated steps (that I'm improving now). I've been reading quite a bit about the money transmitter stuff, but I'
704.
▲
Ask HN: Bitcoin business unexpectedly taking off, help!
152 points
by
mhluongo
13y ago
|
142 comments
705.
▲
by
mhluongo
13y ago
I definitely see your address in the logs- let me know if you have any problems and I can check with Coinbase. For everyone else, it's definitely faster to supply an email and use a Coinbase account, and then transfer from there at you
706.
▲
by
mhluongo
13y ago
Sorry it took so long- I'm hand-verifying transactions to be safe. Very much a human error I plan to remedy.
707.
▲
by
mhluongo
13y ago
Yep, and if you buy them at a discount on eg eBay so much the better.
708.
▲
by
mhluongo
13y ago
Not sure this is quite the same thing. I'm just trying to buy gift cards for BTC.
709.
▲
by
mhluongo
13y ago
Any citations on the FinCen fears?
710.
▲
by
mhluongo
13y ago
Yeah I'd like to try Amazon but the more liquid cards seem to be more fraudulent.
711.
▲
by
mhluongo
13y ago
Nah, the Starbucks choice is explained in the article. It's a remnant from the original project to help indie coffee shops accept Starbucks cards. EDIT - I just read the article you referenced. That scam only applies to registered card
712.
▲
New website lets you trade Starbucks gift cards for Bitcoin
(digitaltrends.com)
51 points
by
mhluongo
13y ago
|
56 comments
713.
▲
by
mhluongo
13y ago
Only sell. Gyft sells Starbucks cards for BTC though, check them out.
714.
▲
by
mhluongo
13y ago
Right now I'm starting with Starbucks cards, but plan to expand in the future if there's demand. I'm also looking into helping smaller coffee shops accept Starbucks cards, though that's longer term.
715.
▲
Show HN: Sell your Starbucks giftcards for Bitcoin
(cardforcoin.com)
4 points
by
mhluongo
13y ago
|
4 comments
716.
▲
by
mhluongo
13y ago
Any chance you guys plan to add link hrefs to CSVs? I'd love to use this now, but I need the href for backlinks and future inference.
717.
▲
by
mhluongo
13y ago
Data models that use deep "JOIN"s are way faster on a graph database. You're right about branching factor- if you always traverse all relations, any database will be slow. In most cases, however, you don't.
718.
▲
by
mhluongo
13y ago
As @espeed said, Gremlin will work (or just Groovy + the Java API). Cypher can handle this as of Neo4j 2.0 using the transactional endpoint. I'm not sure whether the old(er) batch HTTP endpoint kept the writes in one tx- I believe it d
719.
▲
by
mhluongo
13y ago
Typically as your data model complexity ("relatedness") increases, it's more difficult to scale. I'm not sure about anything like CAP, but I do know that in graph-database land we have to remind ourselves that general gr
720.
▲
by
mhluongo
13y ago
Yeah, FlockDB ( https://github.com/twitter/flockdb ) comes to mind. I think Titan ( https://github.com/thinkaurelius/titan ) should / will be able to handle this too.
More ›