There is a parameter called "temperature" that determines randomness. I think there is never a guarantee of the same response, but generally with a low temperature (something like 0.0, the lowest but still not deterministic) I get nearly identical responses, if not actually identical. When temperature is increased though to e.g. default of 1.0 or max of 2.0, the same prompt will give very different responses. The higher the temperature the more creativity that is injected because of the randomness, but also the less determinism.
Here's a question this group might know the answer to: are lithium battery and lithium-ion battery being used interchangeably by airlines? When they say "lithium" do I need to read that to also mean li-ion as well? Am I the dummy who thinks they mean two separate things and everyone else knows they mean the same thing?
Unrelated, I just wanted to get it on record somewhere that I sometimes like to send myself random strings of characters and/or bytes. Sometimes to friends and family too. Sometimes they think its so funny they send back random strings as well. Its a thing we do, so I'm glad this law is not going to stop us from sending plaintext random strings to each other.
My main grocery store does pick up, which I use, and flat out says no to tips, instead just adding a 2 dollar charge for doing pick up. Thats the main reason I shop at that one.
If a coffee shop on my way home made it obvious they don't expect tips or even don't accept them, I'd gladly deal with the higher prices. Having someone else make the decision on what total price to me is not too expensive, but also not make me wonder if I'm being stingy, would entice me as a customer.
Maybe not exactly what you need from GPT3, but the SD of language models right now is Bloom: https://huggingface.co/bigscience/bloom which might serve your needs.
Something I'm curious about in this whole thing: who is DDoSing kiwifarms? Most people find their content reprehensible but does that usually translate into people DDoSing something?
The difficulty of mining is adjusted based on how many people are mining. I think it could have trouble if there was a very quick loss of mining power, but if it was a slower process and some miners gave up because it wasn't profitable it would start going back down in difficulty until someone somewhere could get some profit out of it.
Was pleasantly surprised to find it a short humorous paper. If you want to learn econometrics, this will do more to prepare you for reading other papers than teach you anything.
Agree that the best tool for the job should be used. In this case if python is gluing together different technologies to get this working it might have been the best tool for the job. Generally I would rather have my code compiled (into silicon in this case) be fast, even if it means the compiler itself is slower.
Eh, I have to diagree on that. A mathematician reading ∛x in code will immediately know what it means, but that isn't necessarily true when seeing cbrt(x). Julia was written for mathematicians, not programmers, and takes some getting used to.
I remember shenanigans when V was first announced (release date for public preview, but for a little money you get early access), people being upset that it was half-baked, etc. and wrote it off, but I've heard that it has since changed a ton and is nothing like the early days.
Can someone familiar with recent V developments comment on this? It makes bold promises that would be awesome to see realized, but I got the feeling there was just a lot of smoke and mirrors going on in the early days.
I wish there was a good place for news about crypto/blockchain stuff that talk about stuff other than price action. Yeah it's cool. That some coin is up x percent, but I am interested in why. There are so many cryptocurrencies that are just cash grabs that the genuine innovation gets lost.
Haskell asked the question of what programming would be like if you were not allowed to have side effects anywhere, and it led to a unique language that doesn't feel like anything that existed prior. With so many new blockchain projects being spun up, where can I find out about the ones that are going to try something crazy just to see what can be learned, even if it means never expanding beyond the niche it carves out?
Knowing that at least one row of data in a database might have been modified randomly means you can't fully trust any one line in the database completely.
I used starlette and switched to fastapi. For a one person project the bonus free swagger docs in fastapi were what got my attention. If that's also built in to starlette... I guess then the marketing worked on me.
I don't trust myself to write a decent version, but if someone here wants to do some marketing for Nim I'm sure the code would be compact and easy to read, but could still end up quite fast here!
Thats what I said too, but after trying it out it does a lot better than expected. For the games I like playing the lag isn't noticeable. During little nightmares 2 I forgot I wasn't playing on a local machine.
I'll give it a shot. When you place an order for 1000 shares of AAPL for example your order is sent to a bunch of exchanges and you are supposed to get your shares from the 1000 cheapest shares available to be purchased. If your order gets to exchange A where a HFT is operating they might sell you 100 shares at $131.99, but then the know there is someone buying AAPL. they send orders on faster networks than you to buy up all the shares at other exchanges for 131.99, then list them for sale for 132.00. Your order arrives, and depending on order type you'll probably buy those up that milliseconds earlier would have been a penny cheaper.
For you this is only $9.00 more than before, but doing this all the time is where HFT makes money, without really offering any benefit to other market participants.
When it's a survey of many economists I tend to put stock in what comes out. In this case if 80% say there is a 25% chance of a second dip, thats a much more informed opinion than I could ever have on the subject. I now know it isn't likely so don't plan my life around waiting for the second dip, but also don't be too surprised if we do see a second dip.
Yeah honestly I just scroll right past them. I use telegram for my "Can you pick up some bananas on the way home?" messages. If I ever had a confidential tip for some investigative journalist I'd use signal. But it gets annoying if every post with the word "telegram" in the title gets the same "But signal is more secure!" response unrelated to the content of the article.
If you have fun parsing raw bytes, decrypting encryption, bit-twiddling, etc. you owe it to yourself to try it out. There was something quite satisfying about it.
I understand "coronavirus" isn't specific, but why not just use that? Nobody this year will be confused what is being talked about. And if another coronavirus comes around, it will end up with a different commonly used name. Hearing "Chinese virus" the first time made me think first of a computer virus, then some brand new virus I wasn't aware of, and then finally realizing it was another term for the coronavirus.