huginn

Model collapse is going to be a big deal and it doesn't take too much poisoned content to cause model collapse.

Have found, not will find.

There are so many spam sites with LLM content.

I think 19 reuses of a single rocket is more impressive here TBH

Not even close to the second time. It's happening constantly but is getting missed.

Too many people think LLMs are accurate.

11 years? Nevermind use the laptop for sure haha

You'll probably save money in the long run using a pi.

Here's a white paper explicitly proving:

1. No emergent properties (illusory due to bad measures)
2. Predictable linear progress with model size

https://arxiv.org/abs/2304.15004

The field changes fast, I understand it is hard to keep up

Sure thing: here's a white paper explicitly proving:

1. No emergent properties (illusory due to bad measures)
2. Predictable linear progress with model size

https://arxiv.org/abs/2304.15004

Unless you want to call your predictive text on your keyboard a mind you really can't call an LLM a mind. It is nothing more than a linear progression from that. Mathematically proven to not show any form of emergent behavior.

Friendly reminder that your predictive text, while very compelling, is not alive.

It's not a mind.

In my experience DDG and Google are pretty close to the same quality. IE neither is good.

Doesn't matter if it's a prerequisite