Another day has gone, only one day left. Nassim Taleb’s Real World Institute workshop is about to finish but still here I am to report my key impressions.
The fourth has been the most technical day so far, with lectures on probability, data mining and cyber security. Taleb’s lecture on his own work on the history of wars was the best session of the day to me, but Tom Messina’s storytelling about a few quite relevant cyber crimes woke the nerd in me up.
Wars are not getting any better
Nassim Taleb loves debunking myths his way and he joined Pasquale Cirillo in a quest for the truth on the following statement: “times have changed and we saw a recent drop in violence compared to our past”.
According to their research, most of the historiography today is based on flawed techniques, delivering invalid statistical analysis. Basically the whole concept of a “(gaussian) average violence rate” is too naïve to consider and, if we reshape the analysis in terms that take fat tails into account, we discover that there is no basis to discuss any trend and no scientific basis for narratives about any change in risk of a breakout of new wars.
This is quite a sad result, who wouldn’t love to hear good news on the chances of a new world war?
Deterministic? You better be cautious!
The analysis on war casualties and war interarrival time brought all of us reflecting on the implications of using our past to extrapolate today’s policies.
The Black Swan Problem strikes back: we can prove no variable to be deterministic! We will never either be sure of what probability distribution we are in. We are facing here the same problems we mentioned yesterday about decidability or the recursive paradoxes also known as strange loops: we just can’t!
This is a way big issue because it has impact on a broad range of things affecting our lives, going from research to media, from finance to reliability of lecturers at conferences. It has always struck me hearing people adding numbers to their… opinions, in order to gain authority, when it is perfectly evident that those data are completely invented from scratch! Let alone when the speaker adds decimals up! That makes me crazy. Yesterday, happily, Taleb gave me analytical reference to hit those BS makers with sharper weapons.
All in all, even if not with a bad intention, it often feels like we measure things with a calibre, we then compromise by marking the point with a 1cm-wide chalk and then we end up cutting everything with a 5 Kg axe anyway. It makes no point!
Probing reality bit by bit
That — it is needless to say to those who know me better — made me think of all the effort wasted by software development companies, advertising companies and most of the knowledge workers I know out there. They don’t get that it is way better invest time and money in understanding what kind of risk they’re in, project by project, rather than trying to build a safe way to predict their future with models that will never help them — and never did.
I have always had a soft spot for agile methods, it is not a mystery, but here, once again, I am forced to acknowledge how powerful it is the overall philosophic core of the Manifesto: focusing more on the interaction that exists between all the people working on a project and keeping the weight of any single transaction small, we may cope with the intrinsic complexity of a turbulent environment.
This has become a key point of my idea to shift knowledge work away from contracts negotiation to a prudent and courageous collaboration, from a world based on complicated contracts to another suitable for a complex world.
Poisson distributions, the Lindy Effect, memoryless processes, a third domain put in between mediocristan and extremistan… this post could last forever because so many are the things we talked about during our fourth day together. But I have to go: in 5 hours I’ll have to start my last day at Real World Risk Institute. Bye!
Read the fifth and last part of Jacopo's journey: Read World Risk Institute - Last Day.
Check out the full list of our upcoming training courses: Avanscoperta Workshops.