thund

7 ms·

61.
by thund 2y ago
Different damages, one is on blood vessels, the other on optic nerve. So you could even get both
62.
by thund 2y ago
Different damages, one is on blood vessels, the other on optic nerve
63.
by thund 2y ago
Mirror: https://maps.app.goo.gl/soe6WX87ihvu4E1W9
64.
by thund 2y ago
"OpenAI API" is not a "standard" though. They have no interest in making it a standard, otherwise they would make it too easy to switch AI provider. Anthropic is playing the "open standard" card because they wa
65.
OpenAI Predicted Outputs(community.openai.com)
1 points by thund 2y ago|0 comments
66.
by thund 2y ago
Doesn’t OpenAI embedding model support 8191/8192 tokens? That aside, declaring a winner by token size is misleading. There are more important factors like cross language support and precision for example
67.
by thund 2y ago
The answer though has so much info already that it reduces the need to visit the links.
68.
Google soon to owe $1 googol in fines(vice.com)
3 points by thund 2y ago|0 comments
69.
The First True Smartphone(en.wikipedia.org)
3 points by thund 2y ago|1 comments
70.
The Highest Website(worlds-highest-website.com)
2 points by thund 2y ago|0 comments
71.
by thund 2y ago
nice, but still many things to be improved: - iphone mirroring keeps going into "paused" mode, showing a "resume" button - iphone mirroring doesn't unblock apps that need face Id, e.g. whatsapp says "passcode r
72.
by thund 2y ago
some corps, with custom policies
73.
by thund 2y ago
In case your antivirus is censoring the page: https://archive.is/90XGW
74.
by thund 2y ago
it's starting to show now, about 10 minutes after the issue started
75.
by thund 2y ago
after several minutes the status page is still showing all is fine. For a service like GH, anything more than 30 secs is unacceptable
76.
by thund 2y ago
better even, don’t use never-expiring tokens/credentials that need rotation.
77.
by thund 2y ago
It’s called monitoring and it’s all automated. When something triggers then of course someone is expected to investigate. It’s how you run a service responsibly. All big corps do that and things would be much worse otherwise.
78.
by thund 2y ago
It still takes the goat first, doesn't seem smarter than GPT4. > Suppose I have a cabbage, a goat and a lion, and I need to get them across a river. I have a boat that can only carry myself and a single other item. I am not allowed
79.
by thund 3y ago
from the parent: > TBH I would just recommend using that and all its under-the-radar features (like grammar, and built in function calling support with Functionary). I never heard about Functionary, this comment seems to say “I would use
80.
by thund 3y ago
ok so the answer is no, thanks
81.
by thund 3y ago
does Functionary work with generic models like llama.cpp? had the impression you could only use Functionary GGUFs
82.
by thund 3y ago
Your sources?
83.
by thund 3y ago
We gave up on iCloud for file sync, it’s broken on dozens of devices trying to “optimize” storage even when asked not to. Imagine having 4Tb (size doesn’t matter) mostly empty hard drives and not being “allowed” to keep a file copy offline,
84.
by thund 3y ago
Open platforms can be reviewed and fixed more easily and faster
85.
by thund 3y ago
although usually LLMs don't care, I would have also tried fixing the typo “Francico” and see if Claude acts differently
86.
Custom Kinetic PC [video](youtube.com)
1 points by thund 3y ago|0 comments
87.
by thund 3y ago
the days of coding are going to go :-) from assembly to javascript to AI... where no man has gone before
88.
by thund 3y ago
it's a nice TV, though I prefer the video quality of LG screens. Samsung's tv OS though is better than LG's in general, although it could be much better. Enjoy it :-)
89.
by thund 3y ago
yes, I want to install My apps, eg developed by me, on My devices, and have full access to all hardware feats, without fake obstacles and limitations
90.
by thund 3y ago
maybe this? https://github.com/microsoft/chat-copilot

More ›