In my opinion, this AI race is crucial, and if LLMs are considered the future, then investments should be directed towards neuromorphic chips.

1/ The consensus is that the best way to make a Language Model (LLM) work efficiently is to use self-documenting code.

Additionally, one of the most effective ways to enhance the overall quality of a project is to ensure that the code is self-documenting.

Quite interesting for the near future; I can't wait for AGI to tell us that we need common rules and to share resources fairly.

However, I'm not sure that tech billionaires are excited about sharing their wealth for a better world or about following the laws.

Google AI mode is a free service, and I hope you have functioning hands and a brain.

While there appears to be overwhelming evidence supporting your abilities, that may not be entirely accurate.

Does that mean you're not a fool?

1/ Gemini 3 Pro is awesome, and so is Opus 4.5. Claude CLI is an incredible agent harness with enormous promise, but it is maddeningly inconsistent. It can be genius some days and a complete idiot on others, so I can't trust it completely yet.

open.spotify.com/track/1q1l9q...
Once, men turned their thinking over to machines in the hope that this would set them free. But this only permitted other men with machines to enslave them. / Frank Herbert.

Objectively, I don't understand why people are still paying dollars to learn about LLMs when we have these nine lectures from Stanford, which are a goldmine for anyone wanting to understand LLMs in depth.

Well, no, that's not what I made this post for. Does the guy have to be brought into everything?

1/ Fandom: The New Mainstream.
A look back at some of the biggest fandoms of 2025 through the lens of fan art, fan fiction, and fandom discourse. Cozy dystopia.

What is everyone reading right now? I'm reading The Color Purple and The Odyssey, but I'm also reading The Diary of a Young Girl.