youoy

9 ms·

91.
by youoy 2y ago
There is a difference between keeping score of comments/posts and giving a general feed for everyone vs curating a visual personal feed to maximise addiction...
92.
by youoy 2y ago
The mental gymnastics that people working for Anthropic will have to do to ethically align with this is going to be interesting.
93.
by youoy 2y ago
It has always been easier to refuse to do things when you don't have the option to do them, or it doesn't make any difference, than when you have the option and the financial interests are in place. See for example "Safely al
94.
by youoy 2y ago
That's what one half of HN think. The other half: Artificial brains in the verge of singularity show another sign of approaching consciousness. The chain of thought of process performance is exactly human, showing yet another proof of
95.
by youoy 2y ago
What I meant is that I can imagine cases where some part of the dataset may look like R2 and then colapse to have a spike that looks like R1, so it is not a standard manifold where all of it has the same dimension. Appart from that, these &
96.
by youoy 2y ago
I prefer a formulation closer to the mathematical representation. With "kill", there is not a lot of space for interpretation, that is why it works. Take for example the name "Convolutional Neural Networks". Do you prefe
97.
by youoy 2y ago
If I look at the comments of this post my conclusion is that using "true" and "know" is hiding what the actual result is from a lot of people. So people get stuck on meaningless discussions. Seeing this makes me conclude
98.
by youoy 2y ago
I dream of a world where AI researchers use language in a scientific way. Is "LLMs know" a true sentence in the sense of the article? Is it not? Can LLMs know something? We will never know.
99.
by youoy 2y ago
I think "manifolds" in AI are not the same as actual smooth manifolds. For starters I would not expect them to have locally the same dimension across the whole dataset.
100.
by youoy 2y ago
Thanks for sharing! Very interesting. > The correlation is actually not bad (0.53), but our model is very consistently over-estimating the score at the low end, and underestimating it at the high end. This is surprising; some variation o
101.
by youoy 2y ago
And those that can't teach?
102.
by youoy 2y ago
School is not enough, learning by doing is not enough... The sweet spot lies as always in the middle. Certain things can only be learnt by doing, some things are learnt x100 faster if you learn the theory first. The most valuable thing is l
103.
by youoy 2y ago
You might have been recycling what I did at Google 8 years ago. (Friendly reminder not to trust everything you read on the internet)
104.
by youoy 2y ago
>... everyone realised what a horrible society results from... Has this ever happened? The GenAI thing is here to stay we like it or not, the same way mainstream shitty AI recommendations are here to stay. That does not mean there won&#x
105.
by youoy 2y ago
I would say that the complacency part is identifying good test with good coverage. I agree that writing test is one of the best use cases for LLMs, and it definitely saves engineers a lot of time. But if you follow them to blindly it is eas
106.
by youoy 2y ago
If (stolen) data is available to download ok, that would be the accurate definition of open AI model. But "accurately specified" is not because you would need to trust that the person specifying it is actually honestly doing it. A
107.
by youoy 2y ago
I'm a simple man, I like simple answers: the guy just turned 40.
108.
by youoy 2y ago
Once you have decided to use the geometrical representation of the complex numbers, the justification of the norm of a complex number is Pythagoras's theorem. If you want a more "complex numbers" justification, you can say th
109.
by youoy 2y ago
He does this to justify that we can define a norm over C, because the definitions speaks about vector spaces. For me that section is not necessary, because he justifies the formula for the norm based on the definition of the norm, which in
110.
by youoy 2y ago
We are all running out of time, no one can help us with that. But what I can tell you is that is does not matter. This is not something that you need to understand rationally, but something you need to feel for yourself. If it creates anxie
111.
by youoy 2y ago
> Some drug side effects, like mild nausea or headache, might be considered acceptable trade-offs for therapeutic benefit. But others, like liver failure or birth defects, would be considered unacceptable at any dose. This is particularl
112.
by youoy 2y ago
In causal inference we usually call counterfactual the unobserved outcome of a treatment, and try to understand the difference in the effect of this two treatments. Here they go the other way around, they look for how the variables need to
113.
by youoy 2y ago
The drilling and active learning part reminded me of this very nice article on Bayesian Optimization from Distill publication [0]. They explain it for selecting the hyper parameters for ML models: > In this article, we talk about Bayesia
114.
by youoy 2y ago
> If you were in charge of a large and well funded model, would you rather pay people to find and "cheat" on LLM benchmarks by training on them, or would you pay people to identify benchmarks and make reasonably sure they speci
115.
by youoy 2y ago
On a side note, and since you mentioned Roman Numerals in your other comment, I would say that the representation of I, II and III being different from IV is related to how the human brain processes quantities up to 4 [0]. So it is a simple
116.
by youoy 2y ago
Thanks for sharing! I like to look at this example inside the debate of if mathematics are invented or discovered. > That is how Whitehead and Russell did it in 1910. How would we do it today? A relation between S and T is defined as a s
117.
by youoy 2y ago
Thanks for sharing! The Gaussian distribution never gets old. And nice plot of this: > 100 000 points drawn from a Gaussian distribution and passed through a randomly initialized neural network. It gives you a sense of how complex the fo
118.
by youoy 2y ago
But you can find "delve"
119.
by youoy 2y ago
Run to fail, run for success, run to waste time, run for fun. I don't care. You can live every experience while running, that's the beauty of it.
120.
by youoy 2y ago
Then I completely agree hahaha

More ›