Balaji Prasad
By Balaji Prasad
“We believe in the reasons, because we’ve already made the decision.” ~ Daniel Kahneman
We make decisions every day. But maybe we don’t make decisions at all!
It could be that there is a machine inside us that cranks out decisions that we are not even aware that are being made. And that we simply come in after the fact and take credit for these decisions using our extensive argumentative skills to do so.
So, what is this mysterious machinery?
Machine Learning
Our brains are sophisticated statistical machines. Over time they capture a lot of data about things that happen in and around us. They cull and curate fast and furiously as the terabytes, zettabytes and yottabytes of data plummet towards us from inside, outside and from every imaginable direction. They classify, aggregate and transform the contents from this deluge and make them all somehow fit in a manner that seemingly makes sense. Sense! That’s what we are about: we need things to make sense to us, which is, in other words, making it all fit with whatever we desire to have these things fit in with. And this is where things get complicated.
Our brains may be multi-layered machines in which some parts are changeable and reprogrammable, but some just are not. Some parts are not as much like software as they are like firmware: things that can be potentially changed but at exorbitant cost and Herculean effort, to the point where it may as well be unchangeable. These parts may be what we could think of as emotions: the decisive and unyielding parts of us that strike with the predictability, fury and venom of a king cobra when faced with things that pose discomfort.
And then there is the substratum — the hardware — that came with the machine in its original packing, and which is beyond our ability to directly access, control and change. This layer is our gift from nature: the parts that keep us alive and breathing despite our every attempt to hack, hijack and sabotage it.
So, when we “make” decisions, it may be this fantastic statistical machine that hums silently under the surface making us make all these decisions while we tell stories to ourselves about the decisions we make. Because we’re good at telling stories with the words, logic and language that we have. And therein lies a problem!
Garbage in, Garbage out
Good machine learning needs good data, as any AI expert will testify. But with our penchant for imaginative fabrications and creative characterizations, we create narratives, theories and wild fantasies that bear little resemblance to anything living or dead. And all this garbage coexists with the data that rushes in constantly from outside. There is no clear delineation between the two kinds of things – the garbage, and the data – and so we are at the mercy of this swirling mix of sense and nonsense that we do our best to make sense with.
If we can somehow make our internal curator work more diligently, perhaps we could be more conservative about what we allow entry to, and what we don’t. However, there are so many layers underneath in which our curator lies, and may have been compromised by, that we may not even be able to trust the curator to be less than corrupted by emotion and other things. So maybe we need to go deeper … go back in time to fix the issues that are baked and caked in from the past? That sounds good in theory, but using a corrupt curator to uncorrupt itself is clearly a challenge, and a paradoxical proposition.
So maybe all we can do is to be a bit wary of the various things that we think we “make” decisions about, take a little less credit for them, and take everything with a grain of salt – no, make that bushels of salt!
In the meantime, there is the holiday season that is approaching. So, the next best thing to all these desired repairs is to make a decision that could circumvent all these things: Make a decision to be jolly!
‘Tis decision to be jolly!
Balaji Prasad is an IIT/IIM graduate, a published author, SAT/ACT Online and in-person Coach, and K-12 Math Tutor at NewCranium. balaji.prasad@newcranium.com.
