doc_ick

8 ms·

1.
by doc_ick 6d ago
Ontop of this, including a photo edit history in a photo including the original photo would increase the size of a photo to be completely unusable or unshareable.
2.
by doc_ick 6d ago
Unfortunately theres no thanks to the suppliers of training data. US courts made sure there’s no recompense for them, and likely never will be. Seems similar to eminent domain, but without limitations.
3.
by doc_ick 6d ago
100%
4.
by doc_ick 6d ago
Well you don’t repeat yourself if the llm repeats it for you at the cost of millions of tokens.
5.
by doc_ick 6d ago
Sounds like the advice to newcomers is to not worry about trying to learn a programming language. There already is an llm to program for you, and do it better than you could, so just learn how to talk.
6.
by doc_ick 6d ago
Not for everything. There’ll be a more economic version down the road. Foldable phones are a new tech even Samsung is close to 2k so everyone’s testing with high costs.
7.
by doc_ick 7d ago
Nope, ~31 hours when using the inner screen, ~44 hours when using the external. Either way 30 >15 hours for the iPad mini.
8.
by doc_ick 7d ago
Not right now, but I could see eventually. Apple prices are always a bit higher then regular and this is their flagship, the next versions will reduce this original price.
9.
by doc_ick 7d ago
I’m sure that wasn’t healthy and left some kind of mark.
10.
by doc_ick 7d ago
Now was there a realistic health risk involved in that “primary example”? If only everyone just told the truth instead of buying the rights to it.
11.
by doc_ick 7d ago
From my history with llms, this exaggerated example does hit home a bit too close for comfort.
12.
by doc_ick 7d ago
I think you meant to say “due to correct belief, Apple did kill the mini due to sales”?
13.
by doc_ick 8d ago
Curious, I wonder why Visual Basic is so high up on the list?
14.
by doc_ick 9d ago
Fair enough, I think we are in agreement then. I just greatly dislike the platitudes of OpenAI or llm companies being on a high horse with laws, when something like this happens and bends backwards against general ownership. 1 single person
15.
by doc_ick 10d ago
As you said earlier, OpenAI can do nothing wrong whatsoever with the data/computation they have for their “greater purpose”. We disagree on this point.
16.
by doc_ick 10d ago
I would disagree with the Anubis effectiveness falling off, but the llm bros are throwing life itself into training to defeat any “perceived” challenge.
17.
by doc_ick 10d ago
I guess we can agree to disagree, as I don’t think I can ever 100% trust any unregulated company with infinite data/computation. *regardless of their perceived goal They may say they have the best of intentions, but call me skeptical.
18.
by doc_ick 10d ago
That is a large enough gray area for me that I think it should be more regulated. If llms do truly help “a greater good” then the company, govt, or something should pay or work with the data hosters. Otherwise it’s back to llm company needs
19.
by doc_ick 10d ago
Would still say it’s a token predictor, a fancy one though. I suppose we can agree to disagree.
20.
by doc_ick 11d ago
Sounds like a fun llm that’s “fixing” things.
21.
by doc_ick 11d ago
So for you, everyone should only be good as the worst nation.
22.
by doc_ick 11d ago
Thank you for providing an arxiv! An aside, I finally do appreciate single column format now, makes it easier to convert to epub.
23.
by doc_ick 11d ago
I mean if you want to let llms do everything for you go ahead. Wall-e implications aside, it seems like a great self centric life.
24.
by doc_ick 11d ago
If you ignore history everything is perfect, or at least a controlled slight deviation.
25.
by doc_ick 11d ago
Please feel free to cite sources then, otherwise I see no relevancy from you.
26.
by doc_ick 12d ago
Just skimming through here but I think you have the wrong ideas with llms, I’d recommend Andrew Ngs course (correct me if you’ve already seen it or something similar).
27.
by doc_ick 12d ago
Why isn’t ranking the score of an llm output against what it is “supposedly” predicting?
28.
by doc_ick 12d ago
Just like how reading a math book doesn’t teach you math, why do they make us read anyway? (Sarcasm) if reading a blog post didn’t teach someone how a car works how come it “can” work for next token predictors
29.
by doc_ick 12d ago
Oh that’s ez, for software we know that has bugs, one reason could be the cost of fixing isn’t worth it.
30.
by doc_ick 12d ago
100%

More ›