Ciao.
How many new problems have you run into lately? Problems you’d never encountered before. Ones you don’t even remember anyone else telling you they had to solve.
If the answer is “many more than what I expected”, you’re not alone.
Clients keep getting surprised by how often s**t hits the fan from a new direction (apologies for the colorful metaphor).
This isn’t a passing phase, and we’d better figure out how to solve creative problems with equally creative solutions.
Read below to find out how.
Enjoy.
Avy
Last edition’s top clicked link: apparently lots of people prefer to manage their mailbox from Telegram. Go figure.
Consider a parent deciding how to respond when their nine-year-old says they would rather talk to a chatbot about their day than to their parents or siblings. Before 2023, this conversation did not exist.
There is no canonical parenting book on it, no generation of grandparents to consult, no shared norm to fall back on. It is mundane, a parenting question. It is also a problem with no known solution to it.
We are being faced with so many new problems. We do not realize they are new, because obviously none of us has a comprehensive picture of all problems ever faced by humanity.
We usually assume we are never the first to solve a problem. “Someone else must have already dealt with this.” Yet, we are now in a position where many problems are very specific and there is no widespread solution yet. This applies from mundane decisions (how to ask a colleague to stop sending meeting summaries plagued by AI slop) to critical ones (how a school should respond when a video filmed in a classroom puts a student at the center of online harassment).
I wonder: why are more new problems emerging now? Does it make our work life increasingly more difficult? Are there opportunities we are missing?
More problems are emerging because we have invited them to modern society. I’ll explain:
A freelance designer pricing a project today must reconcile shifting client expectations, the capabilities of generative tools, peer pricing norms, ambiguous IP law on AI-trained outputs, and their own brand positioning — ten years ago most of those were non-existent. Problems are caused by the encounter of several forces. The more forces clashing, the more complex the problem. As information travels faster across the globe and more people have stake in more fields, new problems are born. The number of forces in play multiplies the wickedness of the problem.
We see the problem and we see the damage, but we do not know what transformation we are seeking. A public school adopting an AI policy is not aiming at a known target — “successful integration” could mean students producing better essays with AI help, students never using AI on graded work, or a third state no one has yet named. We do not know anymore what success looks like when we are dealing with a new problem.
Even when we know what success looks like, the solution often leads to new problems. Noise-cancelling headphones solved focus in open offices and produced social isolation, missed conversations, and a coordination problem about how to interrupt someone politely. Classic “fixes that fail”: short-term solutions that create the conditions for longer-term harm (Senge, 1990).
I can hear those of you who are thinking “You can always ask AI”. AI can answer common questions, and many niche ones, but only as long as someone has solved them before. The problems left to be solved are either incredibly complex (AI alignment, climate change, antimicrobial resistance) or brand new, like the parent–chatbot situation above. Large language models are trained on existing text; on truly novel problems their output is either hallucinated or a recombination of adjacent solutions that may or may not transfer.
We are left with a situation in which the problems competing for our attention are disproportionately new, disproportionately wicked, and disproportionately resistant to off-the-shelf answers.
I don’t think that it’s a bad situation. It’s generative. More problems means more innovation, more talent emerging, more connections between people, resources, and cultures.
A few guidelines:
Assume it is new, but never totally new. There are always references to existing resources that can help. The parent–chatbot question is new, but relationship research on TV viewers’ relationships with broadcast personalities is directly relevant, as is attachment theory and the developmental literature on imaginary companions. Cedric Chin notes that “expertise is primarily a matter of organisation… experts rely not on greater analysis or greater information, but better ways of structuring or organising their knowledge”. The reference material exists; the work is in retrieving them and letting new solutions emerge.
Assume it is never completely solvable, because new problems can arise, but it is almost always possible to make progress. That is what we have done with problems over the past millennia. Smallpox was declared eradicated in 1980 after a global vaccination program. Childhood mortality fell from roughly 1 in 2 in 1800 to under 1 in 25 globally by 2020. Neither is “solved” in a final sense, since new pathogens emerge and new mortality risks appear. But progress is tangible and cumulative.
Do not search for solutions in your experience; search in your ideas. Your experience is a weak source of solutions. Once a person has found a working method, they keep applying it to subsequent problems even when a simpler solution is available. Prior experience often blocks new perception. Cross-domain analogy is the workaround. George de Mestral inventing Velcro after examining burdock burrs caught on his dog’s fur is the cliché example. Maintain a constant flux of diverse content, to develop a divergent approach to problem solving. The ideas have to be in the head, and the bridge has to be made explicit.
Count on AI to provide ideas but never full-stack solutions. As noted above, AI is trained on what has already been solved or written. On new problems, it generates plausible but unverified output; on incredibly complex problems, it produces locally coherent but globally fragile answers. Use it to expand the reservoir of ideas (surface analogies, suggest blank spots, reorganize material) but not to deliver the final answer. Paul Graham’s advice for original work is consistent: “Try lots of things, meet lots of people, read lots of books, ask lots of questions” (Graham, 2023). The role is broadening the input, not outsourcing the synthesis.
The upside of having new problems is that the demand for perfection and linearity is likely to diminish. Your colleagues, boss and clients are learning as well that reality is more complex, and chances for perfect performance are closer to zero. We don’t see it because we are also getting used to more sophisticated tools, and we are deluding ourselves that they’ll make our life smoother. They won’t; but we’ll get slowly learn to navigate a world with more interesting problems and more stimulating tools to deal with them.
When firefighters meet a new kind of fire. The Mann Gulch disaster: a perfect example of what happens when teams can’t deal with new problems. Article.
The Cobra Effect. Why good intentions sometimes lead to bad outcomes. Video.
Why solving new problems helps teams work faster. The new solution brings much more than the absence of the problem. Eye-opening. Video.
The damage of outsourcing problem solving to AI. It’s not just about damage to how we think; the solutions are often ridiculous. Article.
An unsolvable problem isn’t a problem we shouldn’t try to solve. What wild problems are and what we can do about them. Video.
If you were forwarded this by a clearly amazing friend, or if you just subscribed — welcome! I recommend you start here.
If you enjoyed what you just read, consider upgrading. Masters of Babel Pro members get access to one-on-one consulting, video deep dives, exclusive resources, and much more.
See you on Friday for the Experiment issue.
Avy
P.S. - I am mostly active on LinkedIn and Substack. The first one for work, the second one for my soul. Let’s connect!
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.