Xelynega

20 ms·

301.
by Xelynega 4y ago
In that case isn't "house curry" similar to "soup of the day"?
302.
by Xelynega 4y ago
It's not really "deal with it" it's "learn the tool to start using it effectively". If someone is having trouble learning how to use a lathe, that doesn't mean lathes are a "horrific mess" and ne
303.
by Xelynega 4y ago
"thinking" and reasoning can be done by toddlers with a dataset a fraction of a fraction of the size that even the simplest language models are trained on. I don't understand this thinking that it's x because it looks
304.
by Xelynega 4y ago
It shouldn't really... Measure of a man was about social issues surrounding agi if we assume a perfect agi exists, but the only thing agi and language models have in common is a marketing department.
305.
by Xelynega 4y ago
> I would start a webtorrent+ipfs+i2p site Libgen?
306.
by Xelynega 4y ago
Do research on the state of the art technology, or a homegrown copy of it that likely doesn't exhibit the same features. It seems like they thought it through to me.
307.
by Xelynega 4y ago
> one thing it can do reliably is repeat a pattern Isn't this something we've built into every modern language(and arguably the entire point of languages)? If you have multiple pieces of code that share code with tweaks(to test
308.
by Xelynega 4y ago
It's a large language model, interpreting and reasoning about the prompt is outside of its capabilities.
309.
by Xelynega 4y ago
Wouldn't questioning those pre-existing notions be the first step towards constructing such a machine? It sounds silly to dismiss the idea time could go backwards because we don't have any machines that work that way, since that t
310.
by Xelynega 4y ago
What does that mean for Microsoft who get a $ for each encode/decode to the corporate speak protocol? And what does it mean for the companies that can't afford the tithe to Microsoft to automatically encode/decode corporate s
311.
by Xelynega 4y ago
You're forgetting the part where Microsoft owns the integration, so your options as a new business are "pay Microsoft to decode the emails people are paying Microsoft to encode" or "spend man hours going through those em
312.
by Xelynega 4y ago
I think the issue is more with people marketing/talking about them as "AI". When I think AI I think of something like Skynet. I would assume something like Skynet would be good at chess, able to generate new text, and synthes
313.
by Xelynega 4y ago
You're adding the implication that it needs to beat the "best" to be passable, if you could pay an agi 1% of the salary of the "best" to get an average quality, is that not accomplishing the goal without moving the
314.
by Xelynega 4y ago
But do they? It seems to me that most of the volume(and people) are on these centralized sites.
315.
by Xelynega 4y ago
Cursive?
316.
by Xelynega 4y ago
Maybe cel shaded is what you're looking for. Basically rounding colours to the nearest pallete color so that lighting looks cartoonish instead of uncanny.
317.
by Xelynega 4y ago
Chatgpt doesn't "learn from every conversation", they trained it once and the output isn't back propagated through the network. Keep in mind that Githubs terms are the same ones they willingly ignored to make copilotm
318.
by Xelynega 4y ago
Why would your firewall allow your ipv6 IoT devices to receive inbound connections from the internet? Whats the difference between "ipv6 Nat" and a firewall when theres not likely to be any address overlap.
319.
by Xelynega 4y ago
The issue is that even after you sanitize the instructions from the data, you have to put it back into one text blob to feed to the LLM. So any sanitization you do will be undone.
320.
by Xelynega 4y ago
Thats part of my concern with all the money being dumped into openAI et al after they productive their generative models. If all the money is in generative models that give "good looking" results, who is going to spend money inves
321.
by Xelynega 4y ago
Except we do that for humans and it's still not really relevant to language models. If I were talking to a python programmer I could assume they knew what a for loop was, so could phrase a question requiring the context of one differen
322.
by Xelynega 4y ago
I fail to see the distinction you're trying to make. Creating the input for the llm to get a specific result is crafting an input in a way that works for the model and it's something we even do between different people(you probabl
323.
by Xelynega 4y ago
Because the problem isn't "chatgpt can write science fiction". The problem is that chatgtp can generate spam that's not easily filterable from genuine science fiction submissions without investing human effort, so the si
324.
by Xelynega 4y ago
> We should be approaching the topic far more carefully It's the major players in a billion dollar industry driving the conversation while also benefitting from the topic being approached as it is. I don't see a way for it to g
325.
by Xelynega 4y ago
Why does a legal system have to "treat AI as different than other tools" also have to "look at animals as legal persons" to be consistent? The concern with the 'AI' is that it's copyrighted content being r
326.
by Xelynega 4y ago
> But there’s no practical difference between Procreate and Stable Diffusion. One is a statistical process from millions of copyrighted inputs while the other isn't? I think that's a pretty important difference between the two
327.
by Xelynega 4y ago
Why would they not accept it when it's functionally the same as "downloading an exe and double clicking" which people seem to like about windows? If they don't want the container, they can typically compile the source to
328.
by Xelynega 4y ago
I think they're making the distinction between special cases and error cases. An empty list could be valid(e.x. a search of users providing no results) so you still need to differentiate between "special cases that need special lo
329.
by Xelynega 4y ago
I don't quite understand. In the case where the function is expecting a list, why would you want to execute the logic for "empty list" on an empty tuple? Wouldn't this just be the developer using the wrong comparison for
330.
by Xelynega 4y ago
Except wouldn't the ideal in this case be to use default parameters? def site(message, port=print): port(greet(message))

More ›