johnthewise

11 ms·

91.
by johnthewise 4y ago
When prompted, chatgpt answers you as if it is a pirate.
92.
by johnthewise 4y ago
You can combine those through apis to chatgpt to arrive at a system that does all that though.
93.
by johnthewise 4y ago
You are not just taking the average, answers would have to be consistent with each other.
94.
by johnthewise 4y ago
next bit prediction, or long-form autocomplete, may lead to general intelligence. http://mattmahoney.net/dc/rationale.html
95.
by johnthewise 4y ago
It's very unlikely. The moves it makes are not always good. Any public facing chess engine API or the API they chose would have been at least 'good'.
96.
by johnthewise 4y ago
in which an evidence for zootipic origin couldn't be found.
97.
by johnthewise 4y ago
If the tool itself is 90% accurate, conditioned on the events that somebody reviews it, it passes test and PRs, it can get to 99.9999% accurate very quickly.
98.
by johnthewise 4y ago
Maybe don't approve code you don't understand? If people just lets in code they don't understand into products and it passes PRs and tests, they might have written dumb code in the first place. I see essentially zero risk in
99.
by johnthewise 4y ago
Agreed, it's a problem right now. But conceptually, there are multiple ways (persistent memory in embedding dbs to reference, coupling it with search engines and data sources) it can get better.
100.
by johnthewise 4y ago
Well, at least know we have a benchmark where we can test it. Imagine it at 20% being accurate, being confident while not knowing anything about the subject etc. Do we have a reasonable argument why it can't be 40% in the next 6 months
101.
by johnthewise 4y ago
>When you really start to prod this and think "Ok, this tool by design is going to produce highly plausible but completely untrue information" do we really think the sensible next step is "Better integrate this very tightl
102.
by johnthewise 4y ago
Because it's possible that they build world models internally[1] and they keep getting better at tasks we think are related to our 'intelligence'. If they get better because they build better world models, isn't that awe
103.
by johnthewise 4y ago
The most amusing aspect is that those who insist on calling AI models "just statistical parrots" have to constantly come up with elaborate explanations each time the models do something impressive. Has there been any instance in
104.
by johnthewise 4y ago
The algorithm itself tries to learn to remove noise. This part of algorithm is undeniably a learning process, evidenced by the fact that it gets better along the training process. It just happens to be that, in order to remove noise, you ne
105.
by johnthewise 4y ago
He is saying they are not unrelated
106.
by johnthewise 4y ago
It's difficult to measure whether something is a net benefit for humanity, and even if it were possible, that shouldn't be a factor in deciding whether or not it should exist. For example, if candy is potentially net negative to s
107.
by johnthewise 4y ago
How do you copyright a style though?
108.
by johnthewise 4y ago
Do you mean total war?
109.
by johnthewise 4y ago
Depends, they have kalibr cruise missiles that can hit pretty accurately but they are expensive and they don't have many. Other than that, it's mostly the artillery doing most of the punching and they aren't accurate and subj
110.
by johnthewise 4y ago
Apple. If you try to get around the restrictions, you may get banned, which means the end of your mobile business, so you don't risk it.
111.
by johnthewise 4y ago
Well, there isn'a direct quote where he said this but it's pretty close. https://www.birgun.net/haber/erdogan-bir-musluman-olarak-nas... this headline is the closest thing in it he says: '... and they ar
112.
by johnthewise 4y ago
Well, the game plan seems to be : 1- Increase fiscal spending 2- Increase wages 3- Give out very cheap credits from governemnt controlled banks at negative rates. and you will get increase in consumer spending(happy people) and employment w
113.
by johnthewise 4y ago
It has risen to over 12B recently[1] and they are covering it all with 'net error omissions' aka money of unknown origins[2]. 1: https://twitter.com/ali_hakan_kara/status/157575953305331302... , orange is
114.
by johnthewise 4y ago
But those who do not agree with it do not have any other custom or tradition which they claim is actually the constitution. Original claim stands on the argument that if there was really a single constitution, it would be trivial to write
115.
by johnthewise 4y ago
I don't quite remember whether it was first used in Vit paper[1], but it's a fairly straight forward idea. You take the patches of an image like they are words in a sentence, reduce the size of the patch(num_of_pixel x num_of_pixe
116.
by johnthewise 4y ago
If you hold eth, you are long eth.
117.
by johnthewise 4y ago
You could encode categorical variables into embeddings, they are more natural to deal that way. https://tech.instacart.com/deep-learning-with-emojis-not-mat...
118.
by johnthewise 4y ago
There are plenty of web dev work around and you don't need 10+ years of experience to do it.
119.
by johnthewise 4y ago
It wouldn't be zero. If you wanted milk but couldn't find it in the store/spent too much, you might just give up on buying it.
120.
by johnthewise 4y ago
well, certainly it's possible to fool yourselves with A/B testing, it doesn't mean you must be fooling yourselves. I've also seen similar results in recommendation settings in mobile gaming, not once but over and over ag

More ›