Hall of craft
The 10 craftsmen Scoreflow’s tier ladder honors. None of them use Scoreflow. None of these cards are user accounts. They’re here because the names aren’t random — each namesake anchors a level band because they exemplify a craft that maps to real Scoreflow skill tracks. Grinding harder won’t get you higher. Practising better will.
b. 1986
Neural nets from scratch. Tesla Autopilot. The most watched AI teacher alive.
"The hot new programming language is English."
Why this tier: Entry tier because Karpathy made the whole field feel learnable. If you're just starting, you're on his shoulders.
Reference character · not a user account
b. 1985
Productised the LLM era. Shipped ChatGPT to a billion people.
"The best products are the ones that feel inevitable in retrospect."
Why this tier: Altman anchors the "shipping" tier. You stop tinkering and start delivering. Your PRs start landing.
Reference character · not a user account
b. 1976
DeepMind. AlphaGo. AlphaFold. Chess prodigy turned Nobel Prize winner.
"If we can solve intelligence, we can solve everything else."
Why this tier: Hassabis tier is research depth — finding the problem worth solving, not just the solution to the given one.
Reference character · not a user account
b. 1960
Convolutional neural networks. Meta FAIR chief scientist. Turing Award 2018.
"Scale alone will not get us to human-level intelligence."
Why this tier: LeCun tier is when you start pushing back on received wisdom. You read the paper and see the flaw.
Reference character · not a user account
b. 1947
Godfather of deep learning. Backpropagation. Turing Award 2018. Nobel Prize in Physics 2024.
"Nobody uses gradient descent to figure out their mortgage payments."
Why this tier: Hinton tier is where you start mattering to the next generation. Your artifacts survive you.
Reference character · not a user account
b. 1969
Linux. Git. Every software engineer alive depends on both.
"Talk is cheap. Show me the code."
Why this tier: Torvalds tier is infrastructure-craftsman mode. You build the thing everyone else builds on top of.
Reference character · not a user account
b. 1938
The Art of Computer Programming. TeX. Big-O notation as we know it.
"Premature optimization is the root of all evil."
Why this tier: Knuth tier is rigor as ritual. Your diffs are surgical. You optimise the right thing at the right time.
Reference character · not a user account
1906–1992
First compiler. COBOL. Coined "bug" and "debugging". Rear Admiral, US Navy.
"The most damaging phrase in the language is: it's always been done this way."
Why this tier: Hopper tier is when the tools you build change how other craftsmen work. Your abstractions ship.
Reference character · not a user account
1815–1852
First published algorithm. Saw the machine as a creative instrument a century early.
"The science of operations has its own abstract truth and value."
Why this tier: Lovelace tier is seeing what the system is actually capable of before anyone else does. Original thought.
Reference character · not a user account
1912–1954
Turing machine. Turing test. Broke Enigma. Foundation of modern CS.
"We can only see a short distance ahead, but we can see plenty there that needs to be done."
Why this tier: Turing tier is the asymptote. You define the field rather than participate in it. Practically unreachable — and that's the point.
Reference character · not a user account
Climb the tiers by practising their craft — not by pretending to be them.
Where do you land? →Manifesto