Hahah summarize this random pile of disconnected sentences with no clear theme or through-line.. the summary is likely to be longer than the content.
3 Likes
I think Iām going to respectfully disagree here, and this comes from a matter of perspective. I imagine your experience differ greatly.
When I first started F-ing around with my A.I. project, I tried desperately to avoid linux. Cause I donāt know the structure / shell commands / lingo. I still struggle nonstop with it. In fact without chatgpt doing itās man-in-the-middle act, I would have long since abandoned the project.
So yeah, itās easier / better IF you know the enviroment. And I already know the counter argument, I should just learn it, and I am by bits and peices, but the A.I. layer is the only thing keeping me in the game to learn.
Maybe think of it this way, you can probably code in python / C(# / ++ / various) / or language of your choice. But you could be far more accurate and specific in assembly language. But who the hell would do that if they could avoid it?
If you think of it, itās kind of the development arc of the sciences. Thereās all sorts of math you can longhand out and get very very specific answers, or you can use the well tested and honed approximation. Logarithms are the perfect example here.
I guess what Iām saying is, yes itās better to know and do directly, but if you aināt got the mental muscle or the time, the short version is the thing for you. And history shows that as technology begets technology, weāll eventually get to a point where the short cut becomes necessary just to avoid being buried by the scale / volume of work to be done.
1 Like
Google also moved in the direction of indexing and removed boolean search functionality in favour of more natural language processing, because it saves them money if they only need to return content that might be vaguely relevant instead of crawling pages for an exact match. This is why when you try to do any real searches, you get flooded with pages which donāt even contain your search terms, or pages which include terms you explicitly asked it to exclude. To restore some of that lost functionality you have to use something like Searx or other second pass search executed on your own hardware.
Google is a general search engine. Itās #1 use is to filter URLs in case of typos, and its secondary use is to try to point you in a vaguely helpful direction. As you hone in on your target youāre going to be matching on dictionaries, documentation, source materials, etc. The best way to use Google is to use it to find better search engines that are more optimised for the category of thing you want to find - engines which utilise more sophisticated knowledge of the meaningful structure of the search domain.
I can use the find command on linux to search all my files, but any scenario in which I have to do so is a problem scenario. I ought to already know where my files are, and most of the time I do, because I have a fairly effective organizational schema. However, there are files which are more difficult to categorise, even utilising symlinking, so I also use file naming and header descriptions to help me keep track of my files. This combined with version control, the ability to filter on creation and edit dates, filter on user permissions, etc are all tools I rely on to navigate my filesystem. A mere search feature occupies a vary narrow window of usability where Iām lost enough not to already know where I put my file but not yet so lost that I donāt need more advanced tools to track down that file.
Not to mention, when you rely on searching rather than structure, you find the files that you remember to look for, sure - but what about all the files you forget youāre not finding because you forgot to search for them? Arranging them in a structure makes them re-discoverable.
What I would like, is not a less structured solution that will leave me even more lost in a vague and nebulous swirling cloud of confusion, but more structure, like utilising the relational database model to chart my data for easy lookup. Supervised learning models could be used to help fill in the gaps of such databases while keeping track of whether or not a given record has been human verified, and unsupervised models could help to suggest patterns that humans may recognise as meaningful and useful organisational schema. The core structure and underlying encoding must however provide a well structured skeleton for such peripheral organs to connect to. Our drives will always be flat sequences of i-nodes under the hood, and must be mapped to some kind of structured pattern in order for any neural network, silicone or flesh, to make sense of those patterns.
1 Like
Anyone familiar with Adrian Tchaikovskyās āChildren of Time?ā The mad dash to put all our eggs into a technology inherently untrustworthy just makes me uncomfortable. And that we are optimizing things for it instead of, well, meatbags like us seems like the sort of thing youād never want to find yourself on the wrong end of. Which throws back to the fall of the Old Empire that precedes the series. I kinda hate a lot of the actual storytelling. But I canāt stop reading 
1 Like
Havenāt you heard of ātrickle down optimizationā? We optimize systems around us, for the benefit of those systems, and somehow.. someway.. it trickles down to us meatbags. If it worked for economics it will definitely work for this!
Step 1) optimize
Step 2) ?
Step 3) profit!
1 Like
My favorite nickname for trickle down economics was āthe golden shower.ā
1 Like
( ļ¼ā½ļ¼ ) Indeed! It worked so well letting the media make up fantasy economic theories that economist have never heard of, so letās just tune in to their corporate sponsored feeds and let them make up fantasy engineering theories as well.
If we give the corporate products šāAI friendsā*š more power and authority, surely theyāll empower the little guys! All you have to do is relinquish any direct control of your own life to the corporā I mean āšAI friendsšā!
You are probably the only AI capable of summarizing this much textā¦
4 Likes
Havana Syndrome - Microwave RF weapon purchased from Russian arms dealer
2 Likes
Perhaps the tinfoil hats were useful for something..
2 Likes
3 Likes
- Yes I play vanilla CyberPunk 2077
- Yes I play CP2077 and use mods
- Cyberpunk? Never heard of it.
0
voters
1 Like
I stopped playing Cyberpunk 2077 when I got fed up with Johnnyā¦
when you did play, did you play with mods?
The game sounded cool, then the reviews started rolling in, and I just figured it wasnāt for me. Kinda sounded like a Grand Theft kinda storyline, which Iām not into.
Also, the dicks sticking out through clothes just screams unfinished mess to me.
2 Likes
No, I was too disgusted by the rudeness of the game. So I didnāt experiment with it.
Also, why are the characters with the cool cybernetics often evil in fiction? As if expanding ourselves wasnāt the most human thing that we can do.
2 Likes
So you probably didnāt get to the part where you can tell him that heās an absolute tool?
Also, lots of things have been fixed. You can get through the main storyline now only experiencing very minor bugs a handful of times (ie collision issues that resolve themselves).
Iām not arguing, itās not a āniceā game, but what are you referring to specifically by rudeness?
Iāve posted it before:
So yeah, Johnny was too much of an asshole for me.
1 Like
My bad, I should have remembered!
Totally Unoffical Random Interest Poll I Havenāt Talked to Amal About:
I would buy an unlocked Spark without VK compatibilty and accept that changing the default keys would void my warranty as they are not recoverable
- Nah
- Depends on the price
- Yuss!
0
voters