heap_perms

Yes, a substance like LSD can alter one's self-evaluation, but that doesn't negate the validity of the experience. LSD and similar substances offer a unique lens, a radical shift in perspective that can unveil new ideas, hidden patterns. If you get a the message, hang up the phone.

In the end, all ideas have to be measured against all other ideas.

But I suppose what you're aiming at here are objective, measurable results in a scientific context? Well, we have a huge hole in the research body of about 45 years, thanks to criminalization (it is very tedious to get a study approved on a schedule 1 drug). However, in recent years, it seems to be opening up. They call it the psychedelic renaissance. Numerous studies have explored both the subjective effects and the applications in mental health. One that strikes out to me involves testing on individuals attempting to quit smoking. After a year, 80% (!) of the participants successfully maintained their cessation from smoking. [1] This is after a single dose!

Long rant, the point is, there are measurable, real-world impacts that these substances can have.

[1] https://pubmed.ncbi.nlm.nih.gov/27441452/

Microdosing LSD does wonders to my thoughts. I'm not sure if one can call it "improved cognition" (what does that even mean)

Very interesting read, is the jupyter notebook available?

Reverse Engineering the (client) Github Copilot: It appears to be sending an enormous prompt, considering the 20 most recent files, from which some snippets are extracted:

> By default, the fixed window Jaccard Matcher is used. This class slices up a given file (from which snippets are to be extracted) into sliding windows of a fixed size. It then computes Jaccard similarity between each window and the reference file (the file you’re typing in).

Code: https://thakkarparth007.github.io/copilot-explorer/codeviz/t...

> I wish it showed some progress in motivating way.

There is a very popular plugin called Heatmap which essentially shows you a github-style graph in anki. Instead of commit frequency it counts reviews.

I like the helix comparision. It indicates progress. Reading this sentence I fully expected the word "circle" be used.

A huge chunk of the people who use the internet do not know what a DNS is, nor do they care. So the block, while technically pointless, is not inconsequential

Interesting. Coming from a progressive european coutry, this sounds outrageously absurd. The idea of enacting laws based on the bible seems not just archaic but almost surreal. Are they stuck in the middle ages?

I too, would be very interestd in this.

I did not know this that's actually very interesting.

The education system is a system of indoctrination of the young. Always has been.

This is seriously disgusting behavior from Microsoft. Windows truly is an abomination. At this point I refuse to work at places where I'm forced to use windows. I'm convinced that this would affect my personal well-being.

> 30 million Americans marched in opposition to the Iraq war

And it was an illegal. George W. Bush is a war criminal.

That was exactly my use case. I'll look into it.

Thanks, I didn't know that. I'll check it out.

Edit: Just tried it, not working: Tabbing on the xpi file -> Open with. Getting the toast "None of your Apps can open this file" even tough I just installed fennec.

Indeed. They have too much authority with the marketplace. Does someone know if it is feasible to build Fenix from source, fork it and then add a startup function that sideloads any extension the user chooses, not just the blessed ones? I'm considerung doing this, but don't want to spend a lot of time doing something impossible.

I want to load my own XPI files.

Because apparently, I'm not allowed to embrace the full potential of my browser by installing any extensions that haven't been hand-delivered and serenaded by the majestic Firefox Marketplace.

Please share if you're familiar with a method. I'm contemplating the creation of my own customized version of Android Firefox, incorporating code to enable installation from XPI files.

Just wondering, in what context did you have to do that work? Seems like an interesting rabbit hole to dive into.

You've listened to Hendrix live? Well now I'm envious ;)

Yes, that talk is legendary. Very impressive how he is able to both talk and write the code at the same time.

> Powered by Metasearch technology, Dogpile returns all the best results from leading search engines including Google and Yahoo!, so you find what you’re looking for faster.

According to their 'About' Page, they still do it right?

This is great. I like the no-nonsense design, I especially like the inline-approach in the video, where it kind of just pops up. I was afraid it would be a link, but turned out to be very smoothly integrated video, only slightly interrupting flow.

That's a fascinating idea. (Also, personally I want full text search.)

> It felt like some 20 year old was explaining how to life a good life with all the experience 20 year olds have.

That hits the nail on the head. I cannot comprehend how this book is so popular. It literally covered in dozens of cute little platitudes like "take little steps". It does not contain a single paragraph of original thought. It's just so... naively simplistic, so _obivous_.

I was thinking something similar. I cringe when I read stuff like this. At this point it's satire.

Yes I was thinking the same thing, use whisper.cpp for better performance and simply grep it.

Right? That would be much more interesting, and I wonder why they didn't show packet data. They already went through the work of setting up wireshark, might as well peek into the packet data to see what's going on.

I sincerely think that even if all drugs were legal, the rate of consumption wouldn't necessarily increase much. "People who take drugs "don't give a damn if they're legal, generally speaking.

That sounds very interesting. Anyone know where I can find more on this story?

I was thinking something similar. Turns out that you don't need all the weights for any given prompt.

> LLaMA 30B appears to be a sparse model. While there's 20GB of weights, depending on your prompt I suppose only a small portion of that needs to be used at evaluation time [...]

Found the answer from the author of this amazing pull request: https://github.com/ggerganov/llama.cpp/discussions/638#discu...