Disqus - Latest Comments for jasonkolb

Re: Getting big data in shape for the development marathon (part 1)

jasonkolb — Sat, 26 Jan 2013 10:54:24 -0000

I think you've hit the nail on the head regarding data quality and readiness for "deployment". There's a lot of work to be done to normalize things from big data from multiple places. However, I also believe that 100% of the time what you ultimately want from unstructured data is "ready-to-go" structured data that tells you about the unstructured stuff it originated from.

Best

Re: The Doctorless Future of Medicine

jasonkolb — Sat, 08 Sep 2012 10:58:15 -0000

Good to hear from you Alex... I'd love to hear your thoughts on big data, your comment was a bit surprising. To me, this is what big data is all about--although I will admit that the meaning is getting fuzzed to a ridiculous level, reminds me of "Web 2.0" in a way.

What are you working on these days? I'd be very interested to hear what your thoughts are about data in general since the Base4 days (did I get that name right? ;)

Re: The Doctorless Future of Medicine

jasonkolb — Sat, 08 Sep 2012 10:56:41 -0000

That's awesome. Not sure how I feel about it being used to push kids like that, but I would love it personally to push myself...

Re: Are (traditional) financial markets broken?

jasonkolb — Wed, 31 Aug 2011 08:02:07 -0000

This is the primary reason I pulled out of the equity and bond markets in the 2008 time frame. I was day-trading and it was plain as day that I was competing against thousands of computers who had a much more direct line into the market than I did.

I became much more interested in (relatively) illiquid markets where the reward for finding a pricing inefficiency is still there, as opposed to the public markets where pricing inefficiencies are not only spotted and acted upon in fractions of a second, but the simultaneous action of hundreds or thousands of computer-driven traders seeing the same thing resulted in a burst of insane volatility.

I don't see a way to fix this problem without either limiting the rate at which computer-driven traders cN get in and out of positions or shutting them off completely.

Re: Geckoboard: “It’s Chartbeat For Everything Else”

jasonkolb — Wed, 01 Jun 2011 17:04:21 -0000

Nice product, the 30 second update rate is kinda 2005'ish though.

Re: Relationship Strength

jasonkolb — Tue, 19 Apr 2011 10:29:03 -0000

That's an interesting idea, actually. Now that I'm thinking about it, that
method coupled with a degrading hash as the sub property could do some
really, really cool things.

Thanks for the comment!

Re: Relationship Strength

jasonkolb — Mon, 18 Apr 2011 23:26:32 -0000

Thanks Ted, I'll check into that. I looked at neo4j a bit and that use case wasn't obvious to me, but I probably just wasn't educated enough in using the graph database structure. Interesting thought.

Re: Data Lenses

jasonkolb — Fri, 18 Mar 2011 11:08:34 -0000

From their site it looks like they're aggregating predictions from experts to create forecasts. Do they let you switch between expert predictions to see the effect on projections?

Re: Data Lenses

jasonkolb — Fri, 18 Mar 2011 11:06:03 -0000

The other bug consideration is that you have to figure out the weighting for each factor as well. Total square feet is much more important than the type of floor, for example. And the weightings will shift depending on which cluster the house in question falls into. It's a hairy problem. But it sounds like you have the idea.

Re: How can I tell if I am failing at my entrepreneurial venture or start-up?

jasonkolb — Wed, 16 Mar 2011 14:02:34 -0000

Loved this post!

Re: You Are Who You Follow

jasonkolb — Thu, 03 Mar 2011 10:53:41 -0000

Some of the most influential people I know abhor the idea of twitter and facebook. Good luck quantifying them.

Re: Going Faster Than Realtime with Brute Force

jasonkolb — Mon, 28 Feb 2011 09:55:26 -0000

That's interesting, I will look into curve matching algos, thanks!

Re: Analytics and Outliers

jasonkolb — Sun, 13 Feb 2011 13:13:51 -0000

Sasha - nice find. They have a really fun spin on time series reporting. I love seeing companies totally throw the old reporting paradigm out the window.

So what it looks like from my browsing is that they parse text from everywhere they can and pull out the dates that it's talking about as well as the subject. That's pretty cool.

Off to listen to the NPR bit on them now, will make a nice listen for a Sunday afternoon :)

Re: Analytics and Outliers

jasonkolb — Sun, 13 Feb 2011 13:08:19 -0000

I think you are confusing measuring against a metric (which, you're right, you can just compare to the std deviation) with looking at the full set of available data and identifying patterns that either identify or exclude the outliers. I understand what you're saying about correlation does not imply causation, the best you can do is get within a certain probability that you are correct. And the more data you pile into that analysis the higher the probability that you're right.

I think what you're looking for re: discovering links and correlations is the semantic web concept, that's pretty much what it's about. I'm not talking about anything that advanced yet, I just need to be able to add as many data sets as I can find and use them as part of my internal analysis.

Re: We are in the AOL days of Social Networking

jasonkolb — Thu, 06 Jan 2011 19:20:44 -0000

I agree, this captures my own thoughts very nicely and succinctly, thanks!

It looks like commenters are getting confused about open vs. closed SOFTWARE instead of open vs. closed SYSTEMS. There's a big difference, and I think there's a strong argument that open systems always win out in the end over closed systems. Even down to nation-states and the way closed societies are slowly disappearing.

Also, I picked up the .atMy.name domain several years back in anticipation of the very federated service system you describe. Ah well, someday when I complete one of my existing projects I'll pick it up if nobody has come up with a decent alternative by then...

Re: A tech question about Twitter

jasonkolb — Thu, 11 Nov 2010 15:13:20 -0000

Interesting article, thanks. It's probably the most thorough treatment of the topic that I've seen, was a good one.

Re: Wave, don't get me started on that. What a botched launch, just unbelievable.

Re: A tech question about Twitter

jasonkolb — Thu, 11 Nov 2010 15:12:29 -0000

Ah, of course, why didn't I see it earlier... piece of cake right? :)

Seriously though, I suspect you're correct, it's the only way I can think of to link the data.

Re: Diaspora Revealed: Sparse, But Clean; Source Code Released

jasonkolb — Wed, 15 Sep 2010 21:14:06 -0000

They need a big player like google or Microsoft to get behind this or it won't get any penetration. Maybe twitter?

Re: This Week in Startups #77 with Tony Haile

jasonkolb — Wed, 15 Sep 2010 16:07:15 -0000

I just found the show and am really enjoying it, great job. Just one small uh... "suggestion", you really need more of a transition between segments. I had it on in the background and when you start a new segment I get totally lost trying to figure out where you just segued to. Just my 2c

Re: Apocalypse and Bubbles

jasonkolb — Sun, 05 Sep 2010 10:35:47 -0000

In contrast with others here I found this essay very thought-provoking and actionable as well. I have been having similar thoughts myself, minus the globalization angle. But the idea that the alternative to a globalized world is a feral one makes more than a little sense given the state of today's technology. And he's right: that idea shuts down many investment philosophies that would otherwise be perfectly rational.

Swing for the fences or die trying is an interesting way to do business.

Re: Apocalypse and Bubbles

jasonkolb — Sun, 05 Sep 2010 10:29:45 -0000

The central banking system is already pretty much a global thing, so I am not following how savings would be the only capital available to invest. You will always be able to create capital via borrowing as long as this banking system exists.

Re: Apocalypse and Bubbles

jasonkolb — Sun, 05 Sep 2010 10:25:41 -0000

The difference is that Peter is arguing that if you lose (and get an apocalypse), the money you would have saved by not gambling is worthless because society will have reverted to a feral condition. I can't help but wonder if he's right.

Re: Startup Architecture

jasonkolb — Fri, 06 Aug 2010 11:36:04 -0000

Thanks for that awesome comment :)

Re: Why did so many successful entrepreneurs and startups come out of PayPal? Answered by Insiders

jasonkolb — Mon, 28 Jun 2010 15:04:00 -0000

This warmed my heart cockles.

Obviously they had bad experiences with the prodigious layers of management at large companies, who spawn meeting-hell, as have I. It also seems that they were successful in putting in place rules that prevented technical masturbation circle-jerks. Kudos to them.

Re: Beyond Blogging

jasonkolb — Tue, 22 Jun 2010 09:01:49 -0000

Interesting, I hadn't heard of squarespace, thanks for the heads up!