Ludwig Boltzmann

Sparse neural networks are the future.

Nature does not hurry, yet everything is accomplished.

A journey of a thousand miles begins with a single step.

[Ludwig Boltzmann @SparseRecNet]

I will start a new and open to the world stock database to keep track of how an AI system can invests with the goal of revisisting the returns every 5 years.

It starts with

10k Kaspi stock, August 2026

[Ludwig Boltzmann @SparseRecNet]

@Nils_Reimers @srchvrs Then the benchmark is simply too difficult to expect those algorithms to work. Going over the url you shared it is stated **Benchmark for users, not research**. I guess it's not a paper.

[Ludwig Boltzmann @SparseRecNet]

@srchvrs @Nils_Reimers @jeremyphoward yes, that's why you would get rich since they can´t prune the search space in high dimensional data. Curse of dimensionality is here to stay.

[Ludwig Boltzmann @SparseRecNet]

@srchvrs @Nils_Reimers @jeremyphoward Tell it does not matter to a kd-tree and get a penny every time it can´t prune a branch ... get rich!

[Ludwig Boltzmann @SparseRecNet]

@Nils_Reimers I am not sure I get the argument: A general algorithm that is supposed to handle arbitrary data can't handle data from a simple distribution so the problem is the people that uses 'easy' data to benchmark a 'general algorithm'?

[Ludwig Boltzmann @SparseRecNet]

@Nils_Reimers I am not sure I get the argument: A general algorithm that is supposed to handle arbitrary data can't handle data from a simple distribution so the problem is the people that uses 'easy' data to benchmark a 'general algorithm'?

[Ludwig Boltzmann @SparseRecNet]

@qdrant_engine @Nils_Reimers @Saboo_Shubham_ @JinaAI_ @elastic @weaviate_io @qdrant_engine Parameters in the benchmark seem fixed though.