Meanwhile, it used to be a standard talking point that girls need more female protagonists to identify with, same with other protected classes.
It would be nice if the arguments were applied consistently.
23,486 karma · joined February 12, 2014
Meanwhile, it used to be a standard talking point that girls need more female protagonists to identify with, same with other protected classes.
It would be nice if the arguments were applied consistently.
Huh. Was assuming that this was just Tonies for Game Boy enthusiasts, but seems it actually reads the songs from the cartridge. Not bad!
But to me both answers to "why not use WHOIS" data and "why not guess the email address" is unsatisfying. Yes, a lot of sites will not be reachable, but a lot might still be. Instead, they apparently didn't even try?
What exactly would be "responsible" about that?
How is that better than just directly publishing the entire list?
(And the other half are Apple vassals)
But sure, let's talk about Humanity.
Don't give the USB consortium any ideas...
We didn't have an equivalent election here yet.
I mean, the data center they were protesting is literally built, operated and owned by Microsoft.
I agree, AI as a technology is here to stay in some form, but wouldn't you be a little pissed if some foreign company takes your data, builds a data center in your community without you having any control, puts you out of a job and then sells your own knowledge back to you?
Doesn't sound very sovereign to me.
(Not to start with the whole angle that the AI goldrush undoes what little progress in fighting climate change was archieved)
Ah yes, "rabble".
And I have no idea what an "OmniCause" is.
(Well, almost. You might have weird floating point issues or other bugs, where the associativity does not apply anymore and knowing the evaluation order would be useful. But I think this would be rare situations)
Ah, that explains it, thanks!
I was wondering why one of the most "liberal left" states, with a strong opposition to the death penalty would have so many death sentences - but it makes sense that the large number of inmates is instead a consequence of them not carrying out the sentence.
As a non-American, could someone explain this to me?
llm -f https://raw.githubusercontent.com/RikiyaT/lm_understand-logs/refs/heads/main/exp1_which_game/runs/pacman/prompt.txt
Very dumb question: The LLM tool doesn't happen to inject the URL into the prompt, and qwen also didn't use any "ps" or similar calls that could reveal the URL?I know this is paranoid, but it gives me a bad feeling if the solution is stored so "close" to the space that's accessible to the model.
Though I suspect if it really had cheated like this, the reasoning trace would be much shorter and maybe not deal with the weights at all.
The association isn't new, see the related slur for gay people.
(The pun in the blog name might also have been a hint)
As usual, some good practical advice (try to understand the societal context your navigating in; don't attempt to fix a problem before you have understood it in depth) which leads to some questionable conclusions (don't try to fix bad situations at all and just learn to accept the injustice because alternatives could be worse?) and then shifts into weird otherworldly stuff where I'm not sure if I'm dealing with scifi literature or things people are actually believing. (darkworld, Agor).
The whole thing has an unpleasant vibe to it that reminds me of the way "pickup artists" used to talk a decade ago.
Also not sure how large-scale ecosystem destruction counts as "Ennui".
There is zero objective necessity for the current data center gold rush, and yet it takes away an enormous amount of resources.
Meanwhile, China somehow manages to turn deserts back into fertile land.
(And also the sharp difference in levels of sophistication you see inside neural nets and outside of it: Pipelines that feed text to LLMs so it can detect highly abstract and nuanced semantic differences in the text's contents - but the pipeline still uses regexes for sentence splitting...)
Company that cautions against the paperclip maximizer already at work building the paperclip-maximizing hardware themselves.
Though I appreciate how this would invalidate the last remaining arguments for the "data centers bring local employment" claim, if the security and construction jobs are automated away as well.
Maybe it will be replaced with a new spin: "Data center? What data center? Oh that one... that's just a wild weed, you can't prove it's one of ours..."
(There is also some irony that "Platonism" has the mention if its creator right in its name, and that "creator's note" has survived the entirety of that timespan)
I think a generalized version of this argument is what "AI bros" are currently trying to sell to the wider internet or even wider society: "Yes, by all means, go on, make your website and put all your creativity and knowledge on there! Why do you care that the site will only be used as raw material for AI outputs and no actual users will get to see it? The users will get the knowledge just fine, the rest is just ego..."
You can think how well this argument works...
Well, is that really so? There is quite a lot of research going on to identify the "meaning" behind certain weights, neurons or directions. I don't see why a human with enough time, perseverance and ability to run experiments with the model shouldn't be able to uncover some of the underlying mechanisms of the model.
But of course not with the speed with which an LLM can do it.
As others have said, while still absolutely insane, I don't find it completely out-of-the-blue surprising. One of the recurring abilities of (current-generation) LLMs seem to be to rapidly grasp very complex systems, where humans would need months or years to reach the same level of expertise.
You can see this in the way a coding agent can "understand" and explain sprawling legacy codebases in seconds.
You can also see it in the projects that are using them to reverse-engineer proprietary IoT protocols or computer games.
But those were all systems where we had detailed knowledge about what structures are there to be reverse-enginered.
If the models can transfer this ability to some other systems where we don't know the underlying structure then I think we will be in for a lot more craziness in the future. (Protein interaction networks? DNA sequences? Ecosystem interactions? Animal languages?)
As for weights, if the models really can analyze the semantics behind specific weights or directions, I have my hopes that they could also give a textual explanation or help designing tools to make "handcrafting" weights more practical in general.