fleischhauf

having worked with half a remote team from Poland which also never turned the camera on I can tell you that faces provide a lot better method of person disambiguation for Germans than polish names. (no disrespect to the Polish language or people, my lack of knowledge therof is to blame here)

you are right you still need the number of clusters

I think the same. it very often has some reasonable suggestion, then I mostly need to fix some minor things and it's good to go. overall it definitely helps to reduce the amount of time required to write code

I wouldn't be so sure about google not being inherently about getting all your data. apple I'm not so sure.

that's the thing, the more people do it the more it gets accepted. the same is happening with drugs for example.

I guess it mus be something like this, I don't think someone would invest a rather large sum without having something like this in place

while I agree, I think being able to do this by themselves is already way way ahead of German administration capabilities. So you need to see this from a positive angle. I for one am jealous.

the reaction is wrong. it shouldnt be "oh no, jobs are being removed" but "nice, less work more automatization, let's make sure we all benefit through less work and not only the rich with more profits"

I wonder about spaceX track record

without any knowledge of the other processes involved in space shuttle software development, I highly doubt that the correlation is that easy

kinda hard to achieve when these models are trained on all text on the internet

everything is subscription based, because companies have a continuous income stream (as opposed to one time large payment). With an continuous income stream you can plan better.

repercussions on the general economy and for the people is what they are worried about

I was under the impression that the opioid crisis was partially due to doctors just prescribing them and claiming they are an easy pain killer that is not really addictive

AMDs deep learning libraries are very bad the last time I checked, nobody uses amd in that space for that reason. Nvidia has a quazi monopoly, that's the main reason for the price difference IMHO.

it also looks dodgy to be honest. Also if you click too much while logged in they detect "suspicious behaviour" and block you from logging in again for a couple of days without warning, saying don't call customer service, they won't help you (very frustrating if you are short on time). They also have the option of maestro for payment (was decommissioned years ago in Europe) and don't accept real credit cards, only debit apparently.

they could have a well building (pun intended) YouTube channel too tho.

this, if you have money to spend in improving your model, more training data is the first thing I'd take a look at

I suppose the underlying modality does not matter too much if it is about learning per se

I think AWS prices scale with hardware price

but ours is better! also we need our own national standards no body else uses!

buying, selling, production and possession (there are still regulations, but these things are now legal in some form)

downside is, that people seem to think it's not smoking and start vaping inside. disgusting.

a degree does not help against cheaping out on safety for maximizing profit. Making the head of a company personally responsible might help.

is tip jars an actual software solutions for tipping? I'm wondering what is most commonly used for tipping websites/content?

how is this different from some library somewhere?

does it still burn a lot of cash? I'm not super following, but since it's x 90% of employees were let go, no9

it's not about increasing their worth, it's being the top five already and doubling it.

you mean like a .. train

if it gets worse, we will be the good guys tho. it's also a matter of perspective. (not disagreeing with the articles points tho)