Comments on: Rise of the machines (decision making when milliseconds matter)

By: phineasb

phineasb — Mon, 01 Mar 2010 20:09:48 +0000

I agree with you and especially like your thoughts on how adoption will be driven by the cost of making a bad decision. It would be interesting to map out an adoption chart based on an assessment of the costs of being wrong.

Could this be the thing that drives medical professionals to finally give up pen and paper data entry?
Phineas Barnes
Principal, First Round Capital
phin@firstround.com
http://www.sneakerheadVC.com
@phineasb

By: Manuel Martin

Manuel Martin — Mon, 01 Mar 2010 19:34:40 +0000

Hi Phin,

Thank you for this thought provoking post. It seems like the success of the real time web largely depends on the success of the intelligent web. As data volumes increase, humans and applications will need help filtering out the noise to focus on what matters most for a particular task or decision. Some could argue that only operational decisions (e.g. credit approvals) can benefit from real time predictive analytics, I believe strategic decisions (i.e. should we invest in this company) will still need human intervention but the outcome can be improved by having the right information sooner. Real time (or near real time) predictive analytics could seem like a no brainer but not all industries will benefit the same. In addition to the number of decisions, a key adoption driver will be the cost of making a wrong decision b/c of not having the right insight at the right time.

Cheers,
-Manuel
www.in2clouds.com

By: phineasb

phineasb — Wed, 17 Feb 2010 08:55:01 +0000

Great points on the other areas that will be impacted. Thanks for adding to this post with the insightful comment. Also agree on your point about Rodger's focus and think he will not be the last to zero in on this area.

By: famolari

famolari — Sat, 13 Feb 2010 03:39:32 +0000

Great post and important topic. Real time processing and decision-making will certainly find application beyond fintech and adtech, though those areas have clear and immediate financial implications and thus will drive lots of early innovation.

Milliseconds definitely matter in high frequency trading, which is why you see many firms trying to compete on physics by co-locating. But that’s not inherently alpha-generating and advantages will be arb’d or regulated away (NASDAQ said they may start mandating the length of networking cables, for instance.)

Real-time could do some very cool stuff in gaming as well. Imagine the kick-ass, augmented reality scenarios people could come up with if they could interpret massive amounts of visual, online and bio data all in real-time.

Beyond all the fascinating mining of social graphs and data exhaust that FRC’s been talking about, there are many other verticals, many not consumer facing, that are being enhanced by closing the decision-making and real-time information processing loop. From health care (vitals monitoring, sensor fusion) to cyber security and terrorism prevention (anomaly & intrusion detection, pattern analysis) to energy (smart grid demand-response, dynamic pricing, load balancing), the applications are endless. In all these areas, collecting, monitoring and interpreting massive amounts of real-time data in-line adds tremendous value.

It’s a good time to be in algorithms, information processing and high-performance computing. I expect to see lots of new companies and investment focus aimed at leveraging the above to serve specific business needs. Roger Ehrenberg’s new “big data” fund is an early and excellent example.

By: Alex Sherstinsky

Alex Sherstinsky — Wed, 10 Feb 2010 06:07:50 +0000

I agree that the trend that Phineas is describing has the potential to change the data processing stack through its many layers, including the underlying infrastructure. In order for this change to happen, there needs to be a sufficiently broad class of applications that drive and focus these developments around markets. Everyone knows that, which is why in most cases we try to err on the side of developing a minimally viable product and get customer validation. That being said, a lot of us — at least Eugene and I — are hoping to get an opportunity — expressed by customer demand — to develop a powerful technology like that.

By: phineasb

phineasb — Mon, 08 Feb 2010 17:17:04 +0000

I remember the original idea and look forward to seeing where the building blocks of that project lead you in the new one
Phineas Barnes
Principal, First Round Capital
215.327.0849
phin@firstround.com
@phineasb

By: eugmandel

eugmandel — Mon, 08 Feb 2010 11:59:25 +0000

Alex and I are trying a hybrid approach. Remember we talked about our Twitter ad network technology? We stopped doing it (long story ), but the tech we developed came very handy for our new project.

By: Most Tweeted Articles by Venture Capital Experts: MrTweet

Most Tweeted Articles by Venture Capital Experts: MrTweet — Mon, 08 Feb 2010 11:47:08 +0000

Your article was most tweeted by Venture Capital experts in the Twitterverse…

Come see other top popular articles surfaced by Venture Capital experts!…

By: phineasb

phineasb — Mon, 08 Feb 2010 04:09:08 +0000

Yes. Filters will be needed. I wonder if the best filters will be machines (search and recommendation engines) or humans (social recommendations)
Phineas Barnes
Principal, First Round Capital
215.327.0849
phin@firstround.com
@phineasb

By: eugmandel

eugmandel — Mon, 08 Feb 2010 02:49:52 +0000

Excellent post! “Humans do not operate in milliseconds”. Also, humans do not operate on the vast amount of data created by the real time web. This does call for machines to be in the stream and make decisions by themselves.
Another class of applications needed is machines processing the stream and making it palatable for human consumption.

By: phineasb

phineasb — Sun, 07 Feb 2010 21:22:54 +0000

Sounds good. Look forward to learning more about what you are up to when the time is right

By: kulsingh

kulsingh — Sun, 07 Feb 2010 21:09:30 +0000

Phin, fantastic post! I agree with your thoughts above and think we will soon move from wisdom of the crowd to wisdom before the crowd (as we have seen in other verticals like financial services via algo trading). The potential applications that can built on this framework will be compelling and exiting. I am working on a new venture (Tickreel) that is working on this infrastructure and is days from a prototype. Our platform hits on many of the themes mentioned above. Great to hear FRC provide insights on their vision, especially since it aligns well with ours. I'll reach out to Josh and Howard when we are ready.

By: uberVU - social comments

uberVU - social comments — Sun, 07 Feb 2010 16:00:35 +0000

Social comments and analytics for this post…

This post was mentioned on Twitter by kentgoldman: RT @phineasb: Blog post: Rise of the machines (decision making when milliseconds matter) | Separate Piece – http://frc.vc/38v...