b'Education matters NK: I leave literature and fine arts to theyou to eliminate the wrong things andnothing but algorithms, and there is no discretion of the people getting thekeep the right things. With revolution,such thing as a universal algorithm. So, education. Still, subjects like philosophy,you destroy everything, and you mayyou have to look at those things from the history, and mainly, history of science,(and will) destroy the positive as well aspoint of view of why you are doing this. the basis of the political economy, arenegative. If you put data in a numerical model and essential not just to have at the generalget the results, you have to understand, cultural level. Its important to disciplineMP: This brings us to the question of theat least in principle, why you got these and organise your brain and to providerole of digitalisation in the clean energyresults and what are the potential this holistic view, and to be able to bring,transition. artifacts that can appear through these in your mind, to bring all the connectionsNK: If we are talking about modelling,results. We trust the computer results between dots that sometimes do notwe should clearly understand what wecompletely. I dont think its feasible.look connected to all. You dont have toare actually modelling, the inputs, and be an expert in philosophy. Still, you havethe parameters of the model. Rubbish in,Another significant aspect of to have at least an initial understandingrubbish out is a well-known fact. Second,digitalisation are databases. The of the fundamental philosophicalyou have to have a robust model. Thatsdatabases are absolutely essential things. concepts and the basic history ofall about balance again. But they are meaningful only if theyre philosophy because this is where all theconvenient to use and have enough science originated from. Such knowledgeflexibility. In the old days, when you gives you a powerful tool and mindsetSome traditions are bad, andlooked through the reference journals to to see your place, the place of whatsome are good, but evolutionfind the right papers, you were looking you do, in the whole picture. Regardingthrough many titles and finding the right literature and fine artthat is a matterallows you to eliminate theones with your eyes. And occasionally, of general culture. Im always advocatingwrong things and keep the rightyou would find interesting things that the development of personality from allthings. you would miss in a modern database angles, but, as I said, its up to a particularsearch. So, if you are making a database persons discretion and interest. Back in the old days, when I was teachingsearch, you have to be very careful analytical chemistry, students gavehow to formulate your search request, MP: My next question is, what should weand its a good practice to formulate it expect, evolution or revolution in tertiaryme the results with six decimal points.several times differently so as not to miss education to ensure a smooth andI told them: You will get an excellentsomething.efficient transition to clean energy? mark, even more than excellent - seven out of five - if you measure this. So,Another problem is that being a NK: I am notoriously against revolutionswhen it comes to modelling, you mustspecialist in your narrow field, you because revolutions typically (a) leadunderstand the accuracy of the result.concentrate on optimising processes to a massive period of disturbance andWhat are the significant numbers?and performance in your area. You may (b) after this disturbance, they relax toFor example, when a forecast of thelook at the performance of a catalyst, another form but of a similar type ofpercentage of hydrogen-based energyfor example. But the performance of system. Therefore, revolution cannotconsumed in 2050 comes with thethe catalyst is not everything. Other change things sustainably. All theprecision of two decimal points, itthings affect the performance of the sustainable changes are happeningcompletely undermines all the credibilityplant. I can give you an example. I evolutionary. of the prediction because it shows a lackwas assigned to the Hydrogen Energy of understanding of the nature of things.California project (one of the first IGCC If you make a revolution in tertiaryOne of the aspects of digitalisation is education, you make massive changes.understanding the level of accuracy. Inplants with CCS). The initial question And here lies the problem: educationmaterial science, theres a good term - fitwas straightforward: What will be the is not a field where results appearfor purpose. right coal milling technology to use? immediately. There is a considerableAnd ideally (at first sight), this process lag time, and if you change thingsMP: In geoscience, we discuss theshould be optimised to minimise energy immediately today, you will see theuncertainty of our results. consumption and CapEx costs. But when outcome of your changes in fivetenwe talked to the people specialising in years, at least. And theres nothing youNK: Another aspect of digitalisation.gasification technology, we realised that can fix. While if you do things slowly andPeople are talking about big data,the particle distribution size (different change one or two items at a time andartificial intelligence, and machinefor various coal milling technologies) observe whats happening, then youlearning - great words. They have tostrongly affects the performance. And have a much better chance of makingunderstand their tools and what theywe ended up choosing technology that adjustments. Yes, it will take longer, butare used for. Why do you need big data?uses more energy to pulverise coal but you can manage the risks associated withBefore doing something, you need toprovides the particle size distribution changes. And all changes are about riskask yourself: Why? Why are we analysingthat increases the efficiency of the management. big data? What exact information do wegasifier. Overall, it was much more want to obtain? Do we want to establishbeneficial than choosing the cheap coal So, I am not advocating any rapidnew relationships between variousmilling option. This was another lesson dramatic changes in the system. It has toparameters? Yes, thats an excellentfor me: you have to try to see the big be done gradually. Another problem withexercise. But what exactly do you wantpicture, understand your part and your revolution, you are losing the continuityto achieve? Why do you want to do thisrole of this thing in the overall picture, of traditions. Some traditions are bad,analysis? I understand that artificialand dont think that what youre doing is and some are good, but evolution allowsintelligence and machine learning arethe cornerstone of everything.41 PREVIEW FEBRUARY 2023'