Predicting the future is a random access process, but inventing the future is a serial process. Most of us (me included) engage with the future the wrong way around: predicting it serially as a narrative, inventing it random-access via spaghetti on the wall pseudo-experiments.
Conversation
Replying to
Prediction: don’t extrapolate the current narrative, with or without discontinuities or forks. It’s fun but useless. Save that for fiction.
Instead, sort small marginal developments into “part of the unevenly distributed future” versus “noise”. Spot seeds, not trees.
1
1
20
Invention: Don’t confuse experimentation with coin-tossing/dice-rolling iid trials.
The power of an experiment is proportional to the length of the serial chain of linked hypotheses it is part of. A 2-link hypothesis chain is 4x more powerful than 2 parallel experiments.
1
3
15
If you experimentally test a first stage and second stage rocket design, you have a launcher to orbit. If you test 2 first stage designs, you have 2 first stage designs. 2 half rockets.
Serialized parlayed bets = create future. Parallel bets = futz around feeling futuristic.
1
2
15
Somehow, agility and iteration have gotten conflated with *small* experiments or worse, *partial* experiments. The risk/return profile of an experiment is a product of the steps in longest chain. You can’t lower risk or increase returns by increasing number of first-steps.
1
1
13
That approach is for people investing in the future rather than trying to invent it. Don’t invent like an investor or invest like an inventor unless you have O(n^2) resources, n=number of hypothesis *chains*
1
9
I have no idea why I’m tweeting this shit. Yet another of my massively parallel idea explorations driven by misguided intuition that 100 tweets will somehow equal 1 blog post 😂
9
