mk67

It lacked Trump's face and branding.

Working at SAP and working language is English, also in Germany.

It definitely will be if you go to court. As soon as you have any witnesses there is little chance to get out of a verbal contract.

No, in basically all countries even verbal contracts are valid and enforceable.

Very nonsensical statement, as it helps a lot since decades and allows faster traffic flow in roundabouts.

Same when I interviewed ~1-2 years ago.

From what I read it's not prosecuted in San Francisco e.g. anymore.

https://en.wikipedia.org/wiki/Principal_component_analysis

Not easy to give a concise answer here, but let me try:

The problem mainly occurs in networks with recurrent connections or very deep architectures. In recurrent architectures this was solved via LSTMs with the signal gates. In very deep networks, e.g. ResNet, this was solved via residual connections, i.e. skip connections over layers. There were also other advances, such as replacing sigmoid activations with the simpler ReLU.

Transformers, which are the main architecture of modern LLMs, are highly parallel without any recurrence, i.e. at any layer you still have access to all the input tokens, whereas in an RNN you process one token at a time. To solve the potential problem due to "deepness" they also utilize skip connections.

I'm in the industry and nobody does that since over ten years. There was just a small phase when Hinton published "Greedy layer-wise training of deep networks" in 2007 and people did it for a few years at most. But already with the rise of LSTMs in the 2010s this wasn't done anymore and now with transformers also not. Would you care to share how you reached your conclusion as it matches 0 of my experience over the last 15 years and we also train large-scale LLMs in our company. There's just not much point to it when gradients don't vanish.

I thought sunlight and water also kills them.

There is no LLM in Tesla cars. Otherwise correct. I'd rather expect they use a convnet or a vision transformer, probably the former.

Adding noise is generally helpful for regularization in ML. Most modern deep learning approaches do this in one way or the other - mostly dropout. It improves generalization capabilities of the model.

In that case the speedup would be linear.

Sounds like bs if you look at Russian numbers.

When I interviewed at IBM in Germany in the late 00s I was rejected and got crystal clear feedback on why - and that is a big American company in Germany.

How would you get the money you possess and want to launder on the stolen credit card? That more sounds like a method to extract money from stolen credit cards, but not a valid way to launder money.

Functions are by definition not random. Randomness would break: "In mathematics, a function from a set X to a set Y assigns to each element of X exactly one element of Y"

Come on, it's pretty delusional to think large scale transformer LMs alone could ever reach AGI.

In my experience the Icelandic vikings are quite a bit bigger on average. ;)

The truth is that quite some jobs pay >=100k in Germany. Usually needs to be one level above senior engineer, so staff eng. or some kind of manager.

Don't think it's necessarily wrong though. I can't think of many professions where a large amount of professionals are pretty bad at their jobs.

It's not possible to count letters for an LLM; it only "sees" tokens.

Yep, playing too much Quake online I hit a record 500 EUR phone bill once (1000 deutsche Mark back then). My parents weren't amused. :D

Nobody used TF-IDF for vector lookups without applying a PCA first though.

Literally quite a lot of people have guns in Europe though, many many millions: https://en.wikipedia.org/wiki/Estimated_number_of_civilian_g...

Nevertheless never heard of such a thing happening.

I must say I never had this train of thought. Whether I'm unhappy or happy I greet my neighbor as it is just common courtesy.

You die anyway at some point. Not sure what this obsession about maximizing lifetime is about. I rather live like I prefer to and when I die I won't care anymore anyway.

Very easy for me, never managed to see or imagine anything when I closed my eyes. It's always black no matter how hard I try.

Seems pretty hard without teleportation as your mouth is closed when doing this mostly.