5 ms·

Greplin (YC W10) Grabs $4 Million From Sequoia For Social Search

138 points by ssclafani 16y ago

wildmXranat 16y ago
I like the idea of searching my own online service accounts, but how do they implement visibility layers ? There are reasons why online services don't allow indexing content. The outcome could be either really good or a privacy nightmare.
- knowtheory 16y ago
  Most of the service revolves around you authenticating greplin (using oauth or xauth) to provide them with a feed of your data. Greplin's core service really is tied into reading all of your stuff. That is a privacy nightmare in some regards, but all of this data exists in other services already, so you trust someone else with it. Question is whether you trust Greplin i suppose!
  - arst 16y ago
    There's also a difference between trusting various bits and pieces of your data to exist on a variety of online services, and trusting a single service to aggregate all of your data from everywhere. A local version of Greplin would be more interesting to me; I think the Locker project https://github.com/quartzjer/Locker https://github.com/quartzjer/Locker might make that possible someday.
troyk 16y ago
Congrats Daniel, sometime would like to know about the backend rewrite
- smanek 16y ago
  It involved a lot of interesting work, including porting all the infrastructure from PHP to Python and Java. We were shooting to make the whole system less tightly coupled so that, for example, the machine fetching my GMail data, the machine indexing the data, and the machine displaying the front-end website are all able to survive and do work if any of the others are down. This allows us to scale horizontally just for our current bottleneck, and to be more robust against individual machine or external service failures. Additionally, it helps us build in much deeper security since each piece of our infrastructure has a small and well defined purpose.
aepstein 16y ago
Not only was Daniel (and his co-founder Shai) the youngest in our YC batch, but they built and launched a couple of different startup ideas in the time that the rest of us built a couple of features. Daniel easily could've packed it up when he lost his cofounder - Shai had to go back to Israel a week or two before Demo Day. Instead, he built a prototype of what would become Greplin (in a week), pitched it to investors at Demo Day, secured funding from Sequoia, and is riding a rocketship. When PG and YC say that they invest first and foremost in founders, this is a perfect example of what they're talking about. Every startup hits a rough-patch, and some even look like they have no hope at all, but the best founders persevere against crazy odds and will themselves to succeed.
- 6ren 16y ago
  So Daniel was a solo founder, for Greplin?
  - danicgross 16y ago
    for a bit. then i met robby walker.
    - billclerico 16y ago
      brilliant pickup by the way. robby is a god.
- maxdemarzi 16y ago
  I never really got the "we invest in founders" idea until now. YC funds people who execute.
- jganetsk 16y ago
  I am going to risk being downvoted to hell on this. Maybe Daniel isn't "persevering against crazy odds" and "willing himself to success". It's apparent from the story that he's a really fucking good coder. With a reasonably accurate estimation of his own abilities, and with all the support and inspiration around him at YC, it didn't take much perseverance and extraordinary will to put something together. He just had to continue doing what he loves and does best... writing code. I certainly can't speak for Daniel. I get tired of these descriptions of entrepreneurs that make them sound like cancer survivors. I think founders often have extraordinary skill, and that changes the required personal characteristics considerably.
ankimal 16y ago
I would love to know what their stack looks like and if they re using any open source libs especially what they re using for indexing and search.
- smanek 16y ago
  Some of the main pieces of our stack include Lucene, Twisted, and Tornado. We're looking to do some more work with Hadoop and HBase coming up too.
6ren 16y ago
Sounds like a great example of launching, then improving: if there's a real market need, users will forgive everything else.
surlyadopter 16y ago
Last week Greplin gave all(?) of it's beta testers a free upgrade to the premium tier of service. That was a nice touch even though I have not yet found a use for their product.
grandalf 16y ago
Anyone who applied to Greplin for a job based on the programming challenge looking for a job? Let me know b/c my company is hiring for two engineering positions in SF.
- vegita1022 16y ago
  contact info?
  - grandalf 16y ago
    email mmmurf at gmail
moomba 16y ago
I never heard of this site until now. Its a great idea. I really wish I thought of this. Not only is it a fun project, but it has the potential to be used by a great many people. I wish the Greplin guys luck with this.

138 points by ssclafani 16y ago

I like the idea of searching my own online service accounts, but how do they implement visibility layers ? There are reasons why online services don't allow indexing content. The outcome could be either really good or a privacy nightmare.

knowtheory 16y ago
Most of the service revolves around you authenticating greplin (using oauth or xauth) to provide them with a feed of your data. Greplin's core service really is tied into reading all of your stuff. That is a privacy nightmare in some regards, but all of this data exists in other services already, so you trust someone else with it. Question is whether you trust Greplin i suppose!
- arst 16y ago
  There's also a difference between trusting various bits and pieces of your data to exist on a variety of online services, and trusting a single service to aggregate all of your data from everywhere. A local version of Greplin would be more interesting to me; I think the Locker project https://github.com/quartzjer/Locker https://github.com/quartzjer/Locker might make that possible someday.

troyk 16y ago

Congrats Daniel, sometime would like to know about the backend rewrite

smanek 16y ago
It involved a lot of interesting work, including porting all the infrastructure from PHP to Python and Java. We were shooting to make the whole system less tightly coupled so that, for example, the machine fetching my GMail data, the machine indexing the data, and the machine displaying the front-end website are all able to survive and do work if any of the others are down. This allows us to scale horizontally just for our current bottleneck, and to be more robust against individual machine or external service failures. Additionally, it helps us build in much deeper security since each piece of our infrastructure has a small and well defined purpose.

aepstein 16y ago

Not only was Daniel (and his co-founder Shai) the youngest in our YC batch, but they built and launched a couple of different startup ideas in the time that the rest of us built a couple of features. Daniel easily could've packed it up when he lost his cofounder - Shai had to go back to Israel a week or two before Demo Day. Instead, he built a prototype of what would become Greplin (in a week), pitched it to investors at Demo Day, secured funding from Sequoia, and is riding a rocketship. When PG and YC say that they invest first and foremost in founders, this is a perfect example of what they're talking about. Every startup hits a rough-patch, and some even look like they have no hope at all, but the best founders persevere against crazy odds and will themselves to succeed.

6ren 16y ago
So Daniel was a solo founder, for Greplin?
- danicgross 16y ago
  for a bit. then i met robby walker.
  - billclerico 16y ago
    brilliant pickup by the way. robby is a god.
maxdemarzi 16y ago
I never really got the "we invest in founders" idea until now. YC funds people who execute.
jganetsk 16y ago
I am going to risk being downvoted to hell on this. Maybe Daniel isn't "persevering against crazy odds" and "willing himself to success". It's apparent from the story that he's a really fucking good coder. With a reasonably accurate estimation of his own abilities, and with all the support and inspiration around him at YC, it didn't take much perseverance and extraordinary will to put something together. He just had to continue doing what he loves and does best... writing code. I certainly can't speak for Daniel. I get tired of these descriptions of entrepreneurs that make them sound like cancer survivors. I think founders often have extraordinary skill, and that changes the required personal characteristics considerably.

ankimal 16y ago

I would love to know what their stack looks like and if they re using any open source libs especially what they re using for indexing and search.

smanek 16y ago
Some of the main pieces of our stack include Lucene, Twisted, and Tornado. We're looking to do some more work with Hadoop and HBase coming up too.

6ren 16y ago

Sounds like a great example of launching, then improving: if there's a real market need, users will forgive everything else.

surlyadopter 16y ago

Last week Greplin gave all(?) of it's beta testers a free upgrade to the premium tier of service. That was a nice touch even though I have not yet found a use for their product.

grandalf 16y ago

Anyone who applied to Greplin for a job based on the programming challenge looking for a job? Let me know b/c my company is hiring for two engineering positions in SF.

vegita1022 16y ago
contact info?
- grandalf 16y ago
  email mmmurf at gmail

moomba 16y ago

I never heard of this site until now. Its a great idea. I really wish I thought of this. Not only is it a fun project, but it has the potential to be used by a great many people. I wish the Greplin guys luck with this.