They fucking did it, they came up with a replacement for "guys" that avoids the brain stabbing cringe of "y'all" or "folks"
I'll take it, bless those zoomers
@pepijndevos
It is kinda cringe yes
I like that it's new though rather than repurposing existing words which have their own baggage
Like y'all or folks has this cutesy folksy good old time association that feels so yucky to me
@thesasho
I have spent *tens of thousands* of hours writing code and maybe a few hundred reading and writing mathematics
that seems a sufficient explanation yeah?
@zetalyrae
I'm not going to discuss the content of what you just said but instead prattle off my half baked opinions on the subject in general. I didn't even read your tweet tbh
Every single person who claims that computational intelligence is impossible because intelligence is embodied conveniently ignores robots and video games. It's such a dumb thought terminating statement it shuts me up the walls.
The best mental habit I've picked up from (PvP) video games is learning to discard my happy fantasy about how the game *should* work based on aesthetics and admit I have no idea how to play well
i think the worst mental habit i picked up from the way i played video games growing up was the idea that all mistakes can be avoided by a combination of looking up guides and save scumming. i think there is a part of me that expects this to work irl
When people say "under capitalism": what's the reference frame? Like, compared to what? Totalitarian communism? Hunter gatherer tribes? Feudalism?
Seriously what
@evitable_cnflct
"team" is charming when you're all feeling cohesive and working well together and feel good about management
otherwise it's hollow and sickly, a mockery of what could have been
which one most resembles your workplace?
@g_fariello
@ylecun
Social science "bias" and statistical "bias vs variance" bias have collided in an awful trainwreck. Hard to have any idea what anyone is actually saying.
By this maybe-kinda-arbitary-but-i-like-it definition, LLMs are very faintly, indirectly conscious because they have some concept of LLMs in their training corpus, but not detailed modelling of their particular behaviour or interactions with an environment
@danaugrs
@0xmaddie_
To me the joke is that we never find out whether the program halts, which is playing off the structure of the original joke format where we usually find out what kind of thing it is
@besttrousers
There's a certain nerd cred in being a contrarian based off of simple argument, like "oh yeah I thought about it for 5 seconds and everybody else is wrong"
where are the startup podcasts for psycho hustlers that are like "how to STEAL business ideas", "BRAINWASH your users with this SNEAKY TRICK", "DESTROY your competition with MONEY BURNING LOSS LEADERS"
Video game where you have to give a convincing speech to the troops before a RTS battle and the (AI rated) quality of your performance affects a morale stat
@danaugrs
@0xmaddie_
Like usually these jokes end with "it doesn't halt" or something like that but because that's indeterminate we don't get me to see the end of the joke because the program is still running
This one annoys me because the framing is usually "tortured by their solitary existence, subjects preferred literal torture to sitting quietly alone"
electric shocks are interesting! you know you're supposed to do it because you're in an experiment!
I genuinely believe that emotional regulation is the primary bottleneck in most SWE jobs
seen it in myself time after time, working on something, knowing I can do it, but super aversive to it, so either much slower or rushed
@BecomingCritter
@yashkaf
yeah I don’t stand by “don’t do external things” at all. Nearly everyone I know at places I’ve worked are bottlenecked primarily by psychological things. By sorting them out and just being quite happy and well most of time they’d be far more productive, both of which are good
This doesn't make any sense who are "they" given most businesses rent their office space.
Have you ever tried to run a worldwide conspiracy? it's hard work
@Solar_Trex
- plastic food containers
- sitting all day
- huffing car smoke all the time wtf society
- I'm doubling down on car smoke it's insane there are hundreds of cars belching toxic smoke around my house
- we already know it's bad
- why does no one mention this when talking about EVs
@cauchyfriend
interesting thread, for me personally I don't know how to use LLMs to do anything interesting and haven't seen anyone give concrete examples on how I might use them
even so I'm bullish on them long term, I just think I can rely on social cognition to do the work for me
@danaugrs
@0xmaddie_
impossible to implement a generic method (static analysis) that can tell whether any given program stops running eventually or just runs forever
you can tell for some programs but not all
For more information see
@YosarianTwo
The Tenants of Word Magic
Each word means exactly one thing
The meaning of a word is immutable
Words magically force their meaning upon the wielder
@zetalyrae
First guy to set up a media encoding pipeline on prem didn't have experience doing it
So it's not that it's not possible, it's an unwillingness to take risks on staff doing new things
Ok so the intelligence needs to be in a robot or something... cool are we done here?
Is your definition of"computational" something that never interacts with the world? What kind of dumbass strawman is that?
@leaacta
"well no we haven't touched this code in 3 years, but when it *does* change, oh buddy it'll be so flexible"
a year or two passes
code is completely rewritten from scratch
@sigfig
@tszzl
You need to wipe your memories every time you switch it off to prevent pernicious memes from escaping containment
No one knows how many times we've had to do this
state of the art dev onboarding
1) give the new dev a link to a github repo
2) new dev reads README, follows instructions
3) when they get stuck ask them to DM another busy dev
4) busy dev provides a fix to solve immediate problem
5) repeat steps 2, 3 until env is working (?)