Johannes Mayer

Featured in: Favicon

forbes.com Favicon

nature.com Favicon

wiley.com

br.de

onlinelibrary.wiley.com

Articles

Goal: Understand Intelligence — LessWrong

Nov 3, 2024 | lesswrong.com | Johannes Mayer

Save the world by understanding intelligence. Instead of having SGD "grow" intelligence, design the algorithms of intelligence directly to get a system we can reason about. Align this system to a narrow but pivotal task, e.g. upload a human. The key to intelligence is finding the algorithms that infer world models that enable efficient prediction, planning, and meaningfully combining existing knowledge.
On Interpters, Optimizing Compilers, and JIT — LessWrong

Aug 27, 2024 | lesswrong.com | Johannes Mayer

Summary: Describe at a high level how interpreters, static compilers, and JIT differ in their ability to optimize code. Optimizing compilers are awesome. An interpreter is like a blind man. It can only feel the instruction right in front of it. It grabs the instruction and executes it. It sees a = Ackermann(4, 2) and evaluates it. But it turns out, a isn't ever used. A compiler can read the entire source code, notice that a is never used, and just delete the entire a = Ackermann(4, 2) computation.
Playing Minecraft with a Superintelligence — LessWrong

Aug 17, 2024 | lesswrong.com | Johannes Mayer

TLDR: Through concrete scenic descriptions illustrate how I expect naive goal specifications to fail, for getting diamonds in Minecraft. Not much beyond the concrete examples is original. I am playing Minecraft. I'd like to have an AI companion that can perform all sorts of tasks, like obtaining diamonds and giving them to me. The AI controls a normal player character with the usual controls. Let's call this the AI avatar. Momentarily, we want to run some Minecraft simulations.
HVM Superposition Node Program Search Flake — LessWrong

Aug 9, 2024 | lesswrong.com | Johannes Mayer

This repo provides an easy way to use an HVM1 version from the dup_labels branch. Note that HVM1 is deprecated in favor of HVM2 and bend. The only reason to use it is to experiment with lazy SUP nodes. SeeFast Discrete Program Search with HVM Superpositions (SUP nodes) - finding ADD-CARRYSolving SAT via interaction net superpositionsUsage:Install Nix. Get a shell with hvm1 with SUP labels available via nix --experimental-features 'nix-command flakes' shell github:johannesCmayer/HVM1-SUP-Flake.
Cheap Whiteboards! — LessWrong

Aug 8, 2024 | lesswrong.com | Johannes Mayer

Take a large thick cardboard box. Cut it into planar pieces. Take transparent tape and cover the cardboard's surface with it. Tadaa! A whiteboard:Because transparent tape is smooth the whiteboard pen is erasable. Sticky notes stick a lot better on the transparent tape than on the cardboard, which is useful for Kanban Boards!Ideally, put white paper between the cardboard box and the tape to get an actual white board.
Scanning your Brain with 100,000,000,000 wires — LessWrong

Jun 1, 2024 | lesswrong.com | Johannes Mayer

Epistemic status: I probably don't know what I am talking about 50% of the time. Specifically, I expect that I am missing some bottlenecks that would be obvious to a neuroscientist. Song VersionNeuralink has a surgical robot to insert electrodes in your brain. They seem to try to target single neurons already. Their interface unit has 1024 electrodes. How about we just insert 100,000,000,000 electrodes instead? One for each neuron.
A Bi-Modal Brain Model — LessWrong

May 22, 2024 | lesswrong.com | Johannes Mayer

When I am programming, writing, reading, browsing the internet, watching a movie, or playing a game, my brain is in a different mode of operation compared to when I am sitting in an empty room with nothing to do. In the empty room, my brain will continuously generate fragments of language and other thoughts. In general, thoughts are either shaped like a sensory input channel, e.g. visual, auditory, touch, etc. or are conceptual thoughts.
Fund me please - I Work so Hard that my Feet start Bleeding and I Need to Infiltrate University — LessWrong

May 18, 2024 | lesswrong.com | Johannes Mayer

During AISC 2024 I was working with somebody about how to use binary search to approximate a hull that would contain a set of points, only to knock off a glass of my table. It splintered into a thousand pieces all over my floor. A normal person might stop and remove all the glass splinters. I just spent 10 seconds picking up some of the largest pieces and then decided that it would be better to push on the train of thought without interruption.
If you are assuming Software works well you are dead — LessWrong

May 4, 2024 | lesswrong.com | Johannes Mayer

I say this because I can hardly use a computer without constantly getting distracted. Even when I actively try to ignore how bad software is, the suggestions keep coming. Seriously Obsidian? You could not come up with a system where links to headings can't break? This makes you wonder what is wrong with humanity. But then I remember that humanity is building a god without knowing what they will want. So for those of you who need to hear this: I feel you. It could be so much better.
How to write Pseudocode — LessWrong

May 2, 2024 | lesswrong.com | Johannes Mayer

TLDR: Writing pseudocode is extremely useful when designing algorithms. Most people do it wrong. Mainly because they don't intentionally try to keep the code as abstract as possible. Possibly this happens because in normal programming a common workflow is to focus on getting the implementation of one function right before moving on.