Zetobal

22 ms·

31.
by Zetobal 3y ago
Because we don't need 12 boxes when 1 box is enough.
32.
by Zetobal 3y ago
Who would have thought that meth(lite) could do that.
33.
by Zetobal 3y ago
openai, Claude, Mistral large and all models that you can infer with ollama have JSON only modes!?
34.
by Zetobal 3y ago
Same problem with Instagram it's insane that so many giant companies have no rate limits in their recovery flows.
35.
by Zetobal 3y ago
this is the way.
36.
by Zetobal 3y ago
DMA made it a lot cheaper in the eu...
37.
by Zetobal 3y ago
there are 4tb sd cards now.... I paid like 200 bucks for an 8mb usb stick 20 something years ago.
38.
by Zetobal 3y ago
Ferrit beads also work quite well.
39.
by Zetobal 3y ago
2005? You are a bit optimistic on how technology traverses. Smartphone were in an enthusiastic niche at the time and that didn't really change until 2009-2010. Just look at the iPhone sale numbers.
40.
by Zetobal 3y ago
There still is a lot of non-coding dna in there.
41.
by Zetobal 3y ago
Yeah, on paper there is in reality there isn't.
42.
by Zetobal 3y ago
People don't want to buy a third of an inch drill, they want a third of an inch hole.
43.
by Zetobal 3y ago
Don't go replicate built your own flow in comfyui and deploy it serverless on runpod. I pay like 10 cent per image for a similiar workflow.
44.
by Zetobal 3y ago
Stop diagnosing people on the internet especially with a complex sickness like bipolar. I am bipolar and your reasoning shows that you only have coffee table knowledge. Edit: HN really is going to shit.
45.
by Zetobal 3y ago
It's the dataset most of the images that were tagged don't have the text shown in the captions. We do a lot of car loras and if you tag the shown text as an example on the numberplate you can prompt/replace it in your prompt
46.
by Zetobal 3y ago
Someone has a burned in square on their tv.
47.
by Zetobal 3y ago
Free speech absolutist™
48.
by Zetobal 3y ago
1. You probably won't hit 100% parity just by prompt engineering especially if you want to do your points 3-4. You would need to train a style LoRa. 2. You can use one of the many ip-adapter models but the best result would probably be
49.
Ask HN: Starting as a techlead in a non tech org: What should I watch out for?
1 points by Zetobal 3y ago|2 comments
50.
by Zetobal 3y ago
The same goes for YouTube... it circumvents the stupid new search altogether.
51.
by Zetobal 3y ago
Pretty much just spam failed login attempts works for almost every big tech corp account.
52.
by Zetobal 3y ago
Probably not the same problem but we have seen a lot of dev accounts being flagged since they started gating code search a few weeks ago.
53.
by Zetobal 3y ago
why would custom nodes need to comply with gpl?
54.
by Zetobal 3y ago
Right years after RED lost it's competitive edge sounds exactly like the thing Nikon would do. The weapon was the last camera with an advantage and it's now 8 years old?
55.
by Zetobal 3y ago
The same goes for webp normal people don't know what to do with them.
56.
by Zetobal 3y ago
Eh... most llms work better if you social engineer them.
57.
by Zetobal 3y ago
That's not a new problem I've met more than one react dev that can't write vanilla js.
58.
by Zetobal 3y ago
The comments on the article are definitely something else.
59.
by Zetobal 3y ago
Or he just dwells on this one thing for the rest of his life like many others did before him...
60.
by Zetobal 3y ago
Even some toys are toys.

More ›