caveman computer

lolloj/Shutterstock

Voices


Evolution does not work quickly. It takes many generations for our genetic code to adapt to changing environments and circumstances. What this means is that our 21st century human genome is still basically the genome of a caveman.

Our genome was well-adapted to the environment of our hunter-gatherer ancestors, because that environment lasted for hundreds of thousands of years. Unfortunately, the 21st century world we live in bears little resemblance to the prehistoric world. Most of the change has occurred in the past few centuries, and ongoing change is only accelerating.

The prehistoric world was a “small-data” world. It was in this world that the human propensity toward confirmation bias first developed. Confirmation bias is the tendency to favor information that confirms our preconceived beliefs — and to ignore data that contradicts those beliefs.

When a caveman walked in the woods and heard a rustling noise behind him, he had very little data on whether it was a leopard or a squirrel making that sound. Cavemen had to infer danger or opportunity from limited data inputs. When they heard a growl or a squeak, it would reinforce a hunch, and they would act accordingly. But our ancestors were mostly inferring things and acting on hunches based on locally accessible information. The small data set they relied on was a product of their five senses and their limited personal experience of the small world in which they wandered.

A caveman’s life was one of actions, not data browsing. Coming to the conclusion that the rustle behind him was a leopard required a very quick decision. The instant he heard a growl, he needed to act. So forming a hunch — and then being extremely open to rapid confirmation or reinforcement of that hunch — was a critical survival tactic.

As a result, confirmation bias was more important to human survival in a small-data world. It became highly ingrained in our genetically controlled brain development.

But ever since the Age of Enlightenment, and certainly since the advent of the Internet, we no longer live in a small-data world. Today we are being overwhelmed by the data around us. We live in a Big Data world.

On every topic of interest, Google puts petabytes of data at our fingertips. Add to that Wikipedia, crowdsourcing and a thousand TV channels. Every minute of every day, email users send about 200 million messages, and Facebook users share nearly 700,000 pieces of content. Our problem is no longer inferring from limited data. It is finding the valuable signal in the vast noise — finding what is important in that huge mountain of raw data.

I’ve come to believe that confirmation bias has become a human liability in our modern Big Data world. Today, if you believe that childhood vaccinations cause autism, there are websites that will substantiate that conclusion. If you believe in ghosts or extrasensory perception, there are plenty of websites that will confirm and reinforce that view. You can find support for almost any crazy belief, even if the vast majority of the information in an unbiased search might contradict your preconceived notions. Humans are good at ignoring stuff they don’t agree with.

This is an unforeseen consequence of the Internet and the sudden accessibility of almost infinite, diverse data that can be freely posted by almost anyone. I think it partly explains the polarization of our politics, and even the rise of extremism around the world. Now one can find external support for almost any preconceived belief. Different groups can exist in alternate realities that are internally consistent and self-supporting, but totally contradictory to the beliefs of other groups.

Confirmation bias has gone from being a useful survival tool to a source of increasing polarization and disharmony. And I submit this is being driven by our move from a small-data world to a Big Data world. Almost every wild thought can now be confirmed on the infinite Internet.

Once we inferred to survive, now we self-reinforce to disagree. In a strange way, we’ve gone from one reality to many realities, each internally reinforced. This is not a good trend.

Ted Driscoll, PhD, is the partner leading the Digital Healthcare team at Claremont Creek Ventures. His investment portfolio includes AlterG (formerly Tibion), Assurex Health, Cureus, Fluxion Biosciences, GeneWeave Biosciences, GigaGen, Natera, NuMedii Inc. and Zipline Medical. Reach him @easydjr.



10 comments
enki
enki

The premise of this article is that the prehistoric world was a "small data" world. I see no way of coming to this premise without disregarding the ferocious complexity of the world while also ignoring the decisecond data flow a human (or any situationally aware mammalian social animal) filters, interprets and processes.


BTW how is Mr. Driscoll defining small data? I inferred Mr. Driscoll was saying the source of the data was the natural world, perceived stars, heavenly bodies and the human condition etc. Does Big Data include source data?  Would an information theorist agree with Mr. Driscoll contention that a human handles only “small data”? You can have efficient algorithm but that doesn’t make the source data small.


A thought experiment- Given that a robot would have to process and respond to all the same information provided by the five human senses concurrently (in the range of each human sensitivity)- how much total data would have to be processed for a Robot- lets call him Alan- to behave as well as a human in prehistoric conditions for the robot to:

1)  hunt and gather, skin and cook an antelope or sea lion. [I suspect a learning algorithm will be needed for this. Let’s presume that no lions nor other predators will have to be avoided. ]

2) Robot-Alan should also build a controlled fire (without being consumed by it ;).

3) create and repair tools out of whatever materials are available, 

3) invent musical instruments [In fairness Alan will not have to play the instrument.. well ;] and

4) anticipate inclement weather conditions.

5) And let’s not forget Alan has to create a spoken or behavioral language while-

6 -Optional ) protecting itself from bacteria and viruses by monitoring and responding to microbial fauna. (My limited understanding is that the body is engaged in biowarfare at all times..and winning.) 

7) Alan should function effectively in climates ranging from latitudes N66th to S66th.

8) Lastly Alan will have to adapt to variable climate conditions and prey behavior.


How much data is that? Perhaps someone at Darpa could give us an estimate. I submit to you that it may not be small. 


The fact that humans do not process information quantitatively does not imply nor suggest that there is less information to initially process. What is the resolution of the eye? I do concede that the data that the human mind processes is less than the data it initially receives. Oh i forgot to mention, Alan has to also monitor and maintain its own condition while uniquely identifying every human face it will ever encounter and interpret all human expression and social behaviors in the context of a shared or unshared experience while incorporating all past known experiences..including the interpretation of say two spoken languages.

RachelSklar
RachelSklar

 "Unfortunately, the 21st century world we live in bears little resemblance to the prehistoric world?" Yes! It's really a shame about our modern world of indoor plumbing, hugely impactful public health initiatives like vaccines (buh-bye, smallpox), and ways to send information - and people - vast distances in the comparative blink of an eye. Come on. This article is truly insane, to somehow imply that we are at a DISADVANTAGE by having access to so much human learning and raw information. (Though some information in support of his thesis - actual studies rather than idle  postulations - might have been nice. It's more than a bit ironic that the one link he includes is to - wait for it - Wikipedia.)  


He never particularly had me, but he lost me at this: "But ever since the Age of Enlightenment, and certainly since the advent of the Internet, we no longer live in a small-data world." WOW. I'm glad this guy wasn't in charge of things when Galileo was around.

Nathan Schor
Nathan Schor

Great articles elicit great conversations in the comments, as evidenced in this post.

options33
options33

What a great article and synopsis of the far reaching effects of big data, on business and indeed humanity. Some of the quotes are extraordinary and highlights the calibration of data and technology, that brings us to the world we live in and will be faced with.



Angus Matheson
Angus Matheson

I totally agree with the premise we are low data creatures living in a big data world & that our brains are not designed for/cannot handle big data. I however totally disagree with the conclusions. 1) It is because we cannot handle big data that we have computers. As the last comment stated they analyze and filter the data for us. It is just like other tools. We cannot hack down a tree with our hands, so we have saws. We cannot walk hundreds of miss in a day so we have cars and plans. Computers are the machines for our minds. And this is really important. Sure big data is used to tell me what to buy on Amazon and what to watch on Netflix. There are big questions that hopefully will be answered too, like why some people get MS and how to prevent a rolling blackout on very hot days. 2) Big data is not causing small mindedness and predjuce. Confirmation bias is human nature. It has existed when we had only the least data. Have people have always clustered with like minded people. When was the last time you saw someone go both to the catholic church and the synagogue? And we also tend to cluster with political orientation. Yes the fringe can find like minded. But for the first time in history you can discover ideas that neither you nor anyone you knew existed without leaving your home. The internet will not get rid of narrow thinking but neither is it fair to blame narrow thinking on the Internet and big data.

Piplzchoice
Piplzchoice

Clay Sharkie said "there is no information overflow, there is filters failure". Mere volume and velocity of data does not constitute "big data", but multiplicity of data sources and data formats does. From that perspective the term "big data" describes data aggregated from multiple departments and multiple data bases (i.e. data warehouse model), linked with data from sources external to a company, in a structured and/or unstructured format. Mining such set of data may produce very valuable intelligence. However, all can also result in waste of money, efforts and opportunities. 

maddiegrant
maddiegrant

Interesting post, but there is Facebook data research (see https://www.facebook.com/notes/facebook-data-team/rethinking-information-diversity-in-networks/10150503499618859) that shows the "echo chamber" is not as bad as we think. This relates to your point that people find data to substantiate their (bad) theories - I agree that is true, but research shows that people are equally likely to see information that counters their own opinions.  


Of course, on Facebook specifically one can still surround oneself with the groupthink...

Odyssey
Odyssey

This was an interesting article, thank you. The author's conclusions seem rather negative, though. This smacks of the "nothing good can come from this" argument that came up over Galileo's telescope, Gutenberg's press, or the first camp fire. 


What if your wild thought is the regime you are living under is oppressive and unjust, and you cannot take it anymore? Find enough other people who share your idea and change can be brought about. Confirmation bias is not solely a one-way ticket to ignorance, it can also be a way of rapidly prototyping ways of survival, even in the modern world.


Evolution is not the same thing as adaptation. The genome may evolve slowly, but we have the ability to adapt very quickly to changing circumstances. We may still have the biological makeup of cave dwellers, but we no longer dwell in caves. Preconceived beliefs have changed as we have adapted in the face of new ideas, wild thoughts are tried and discarded. The earth is no longer flat. Mostly. 


As the flow of information evolves, so too will our uses of it. As experience and evidence grow and merge, there will be less places for unsubstantiated arguments to hide and fester. You could argue that the problem now is we do not have enough data at our disposal. Decision-making will certainly evolve when it is within our power to know immediately if our information is incomplete or incorrect. 


The Internet and Big Data are hammers. You can use a hammer to build a ship, or you can use it to smash your thumb. If you believe Big Data will drown humanity in a Matrix-style nightmare, you can find websites to support that, too. 


This is a time of turmoil and change, as it has always been. For good and bad, barriers  to communication are coming down. And as we venture forth into the great and growing ocean of thought, it is time to learn how to navigate the vessel, not jump over the sides in despair. 


Ted Driscoll
Ted Driscoll

@Angus Matheson  I agree with much of your comment.  But I think we are capable of handling more data than was originally available to us.  I was mostly worried about the impact of big-data on reinforcing closed-mindedness.


Ted (@easydjr)

Ted Driscoll
Ted Driscoll

@Odyssey  Sorry if this came off as overly negative.  That was not my intention. I am certainly aware of the great benefits of the easy access to enormous amounts of information that we have on the Internet.  I was just trying to point of an unforeseen consequence.  Nothing comes for free.  In this case, we need to be open to all of that information and open-minded, not just to the information that confirms our preconceived beliefs.


Ted (@easydjr)

Follow

Get every new post delivered to your Inbox.

Join 300,350 other followers