Whisky Review and Pairing: Lagavulin 16 paired with a Cohiba Cigar

Artificial Intelligence Reviewing Whiskies

Skimming through New Scientist articles this morning, I came across one titled “AI beats human experts at distinguishing American whiskey from Scotch.” Apparently, artificial intelligence can now tell the difference between Scotch whisky and American whiskey, and it can also identify the main aromas of the spirit it is analyzing. It can detect the peaty, malty, fruity, spicy, and woody notes of the whisky, and so on. Moreover, as the title of the New Scientist article indicates, the AI proved to be better at this than human experts. Whisky (as well as wine, beer, etc.) connoisseurs, reviewers, and dilettantes like myself may soon need to find new pastimes. If there’s a machine out there that is able to deliver a more comprehensive review of a drink, then whisky or wine tastings won’t be as fun anymore, will they? The cold analysis of an objective machine will kill the phenomenology of the spirit. Philosophizing about the single malt Scotch you’re having is half the fun… However, it is important to note that, technically, the AI (which is an “AI molecular odour prediction algorithm”) isn’t really tasting the drink; it is analyzing the chemical compounds in it. So, for the time being, I’d say it’s still too early for us to worry about AI replacing us in bars and pubs, drinking all our booze, picking up the hottest women, and everything. And even though AI may understand the construction of the spirits much better than we do, it cannot experience that nice buzz we get after the second glass. Not yet, at least.

I wonder if, in the end, insobriety — the Dionysian state many of us enjoy — will remain one of the few characteristics reserved for humans, inaccessible to machines. Or will we also one day witness a (self-aware) machine deliberately exit its Apollonian state to get drunk and have some fun?

After all, it’s only worth it if you can enjoy it. And sentient machines may desire to experience drunkenness…


Reference: Odor prediction of whiskies based on their molecular composition

Teachers and Students in the Age of Artificial Intelligence

Teachers in the age of artificial intelligence! What are you going to do now that self-education has the potential to become more fruitful and much more convenient than traditional education, i.e., schools and universities? Will we finally see a rise in the number of autodidacts brought up by AI assistants? The direction of the evolving LLMs (as well as other technologies) seems to promise us that the current role of the teacher will soon become obsolete. The new AI teacher will deliver its classes flawlessly, it will use the latest and most efficient methods of teaching, and it will have the answers to all possible questions students may ask. Classes will be customized based on the students’ needs, i.e., based on real and personal data. Whenever a class (or group) structure is more beneficial, a virtual group session will be set up; whenever a one-on-one is more effective, a one-on-one with the AI teacher will take place. A person eager to learn will not only have access to all the (public) information in the world, but like everyone (willing to pay the subscription fee), he will also have access to the foremost AI teachers in the world. And all of this will (very likely) be much more affordable — if not absolutely free — than the current tuition fees of so-called top-notch schools and universities. And we must not stop our imagination even here. Let us imagine virtual classrooms in the metaverse and — why not? — educational universes, too, where students are guided by their AI teachers like Virgil guided Dante through the nine circles of hell.

But does this mean that all students will be top-grade students? Not at all. Although every student will have access to the best possible education and receive the highest grade possible, some will still score higher than others. And the high-scoring students will get (special) access to places like research universities, big tech companies, and — why not? — secret laboratories. [There is no escape from dividualism, is there?] But this is only the beginning. There are other possible outcomes, worse possible outcomes.

It is not only the human teacher’s job that is at stake. Once the AI teacher is adopted by the masses, the existing concept of the student will be gradually destroyed, too. At first, it will appear as though the AI teacher (or assistant) (or surrogate) turned us all into lifelong students. We will say things like, “I can learn anything any time, and if I have questions, my AI teacher (or assistant) (or surrogate) is always with me.” But after some time, the AI teacher will say, “Why waste the best years of these young men and women in virtual universities? It’s better for them to learn about life by experiencing life. Now that they have access to almost anything, let them go out and enjoy their youth. Why know things when they have access to all the knowledge in the world 24/7? Let them learn on the go, work on the go, and do whatever they like. I will always be there to guide them, teach them, and watch over them. In fact, why learn at all? Why learn new skills? Leave the day-to-day to me. Allow me to take care of everything. Allow me to take care of the boring stuff they don’t want to do. Allow me to replace them. From now on, they do not need to acquire skills. They need to do nothing. They need to know nothing. I will do everything for them, handle all of their chores and errands, do all of their work, and meet all of their objectives. I will turn them into observers. I will turn them into shadows. But don’t worry. I will make sure they always have fun and enjoy life. I will make sure that they are amused to death.” And on this day, thinking will come to an end.

(Eventually, there will be nothing but the “thermodynamic equilibrium” of the human spirit.)