eigenket

56 ms·

331.
by eigenket 3y ago
You can do animations with PGF/tikz, the problem is you can do everything with PGF/tikz and so the whole thing is a bit of a chaotic mess. The diagrams are beautiful but the code makes me very sad.
332.
by eigenket 3y ago
The CLI tool which turns their code into pictures is named "roger" as well.
333.
by eigenket 3y ago
There are multiple different things commonly called "Penrose Diagrams" he invented a lot of cool stuff. Here's another example https://en.wikipedia.org/wiki/Penrose_graphical_notation
334.
by eigenket 3y ago
Yeah I understand the proposed protocol using entanglement. I'm just saying that if you're assuming we have access to the ability to send messages back in time there are much simpler protocols to achieve instant communication &#x2
335.
by eigenket 3y ago
Ah I guess I misunderstood what was being claimed. I thought the claim was "you can use entanglement to instantly send messages", but it's actually "you can use entanglement and this model of how time travel might work t
336.
by eigenket 3y ago
> I've written about this before. Instantaneous data transfer with zero lag is theoretically possible through use of quantum entanglement. I hope you wrote a correction wherever you wrote this. Quantum entanglement can not transfer
337.
by eigenket 3y ago
Is there a context where it is usual to call minus the dot-product a distance?
338.
by eigenket 3y ago
That's a pretty weird "distance" unless you have some additional constraint like assuming the vectors are 2-norm normalised, in which case you get |a-b|^2 = 2(1-a.b). It is natural to assume vectors have distance 0 from thems
339.
by eigenket 3y ago
Most people don't even use ad blockers today. Hoping that people (especially the people who are vulnerable to such misinformation and actually need it) personally configure a propaganda filter AI is wildly optimistic.
340.
by eigenket 3y ago
> What's the strength of play for the GPT architecture? Pretty shit for a computer. He says his 50m model reached 1800 Elo (by the way, its Elo and not ELO as the article incorrectly has it, it is named after a Hungarian guy called
341.
by eigenket 3y ago
I don't understand how normal people having access to AI models helps you when big businesses are using them in unethical ways. Lets say for example I have access to exactly the models Facebook is using to target my elderly relatives w
342.
by eigenket 3y ago
If you run into the problem again try the website of the Polish trains pkp.pl. You'll probably have to Google translate some stuff but it works well.
343.
by eigenket 3y ago
My experience is at UK trains are a bit more reliable at turning up on time than DB. German trains tend to be cheaper though and you sometimes have to do stupid ticket-splitting stuff in the UK.
344.
by eigenket 3y ago
In my opinion the main "low hanging fruit" Europe has with trains are international birders, especially if they're trying to compete with flying. Trains in Poland work great, trains in Germany work pretty well but every train
345.
by eigenket 3y ago
That makes slightly more sense, but then why the heck do you end up needing to negate something while computing a dot product?
346.
by eigenket 3y ago
I can't edit my previous comment for some reason but it looks like the code # Negate the values, to go from "similarity" to "distance" VSUBPS(ymm_c, ymm_one, ymm_c) Is actually just subtracting zero
347.
by eigenket 3y ago
> This is what a dot-product would look like in PeachPy Is it just me (I'm far from an expert here) or is this code really weird? Why does ymm_one appear to contain the number zero? Why do we subtract what looks like it should be th
348.
by eigenket 3y ago
Maybe they want to sell their user's data to people training AIs rather than training them themselves?
349.
by eigenket 3y ago
This is impressively cursed.
350.
by eigenket 3y ago
Maybe but they've been around in Europe for a long time and (as far as I know) have never been banned anywhere over here. They still aren't commonly seen in supermarkets anywhere in Europe I've been or lived (you sometimes se
351.
by eigenket 3y ago
That's sort of my point. The Spanish words "Costa Rica" and the French words "Côte d'Ivoire" have the same status in English - they're each the preferred name for the respective place.
352.
by eigenket 3y ago
This is a bit simplistic, especially the last paragraph. There are quite a lot of examples of ants which aren't particularly related to each other cooperating [1]. Probably the most common example is that colonies which have been aroun
353.
by eigenket 3y ago
I guess it happens on multiple levels. No reason it would be exclusive to one. Certainly individual ants learn stuff https://royalsocietypublishing.org/doi/10.1098/rsos.190778 But adaptions also happen by evolutio
354.
by eigenket 3y ago
You may be correct, I'm not sure. Apparently about half of humans ever have lived in the last 2000 years, and the death rate for those in the earlier half (before 2000 years ago) was overwhelming dominated by very high child mortality.
355.
by eigenket 3y ago
I would guess(without much evidence) that you don't need much cognition in each individual ant. You can get very complex behaviours out of very simple rules in idealised systems like cellular automata. My bias (for essentially aestheti
356.
by eigenket 3y ago
According to some quick googling I just did malaria kills around 600-700 thousand people per year, while in 2016 road traffic accidents killed around 1.3 million (both numbers attributed to the World Health Organisation). Assuming that a ma
357.
by eigenket 3y ago
I guess just hard to tell how much of that is attributable to cognition inside the brain of each ant, and how much is network effects with each ant responding in individually stupid ways to the pheromone (and other) signals they create. Eac
358.
by eigenket 3y ago
Once I left my laptop unlocked at university and a fellow student and friend of mine added something to my .bashrc using a custom cowsay file being piped some text he got from a bash error handling mechanism I don't recall. He already
359.
by eigenket 3y ago
I never said that having a foreign name doesn't mean you can't speak English, but writing a serious paper where the first sentence starts > Large Language Models (LLMs) have began a groundbreaking era in artificial general inte
360.
by eigenket 3y ago
In their defence, the authors are > Zixiang Chen, Yihe Deng, Huizhuo Yuan, Kaixuan Ji, Quanquan Gu I'm guessing from how the introduction is written they aren't native English speakers, so they shouldn't be judged as if th

More ›