dkalola

Yes, that's right. This is not a practical approach given current LLM limitations. But perhaps in the future if the cost of each individual call decreases, and if latency decreases (and models run locally), but context window remains limited this approach could be more useful as it uses pairwise comparisons rather than loading the entire context window.

How can we interact with such models? Is there a web application interface?

"Systems marketed for "solving crimes" get used for immigration enforcement"

What immigration enforcement are you speaking of here? Legal? Illegal? If the latter, wouldn't this system be solving crime?

Why call it the Trump card? Regardless of what side you are on, this country is bigger than one man. This branding seems flippant and unserious.

This essay by Huxley said exactly what I have always felt but never articulated. The problem is far worse today than it was in his time. So much of recreation today consists of binge-watching Netflix. In addition, because Netflix has global reach, the world often watches the same content, devoid of local meaning.

I have realized that active recreation, although requiring overcoming a certain threshold of activation energy, is far more compelling and fulfilling than passive recreation.

The new video is pretty funny!