Septimaeus

@Septimaeus@infosec.pub

• ⁨Comment⁩ on ⁨Cutting-edge research shows language is not the same as intelligence. The entire AI bubble is built on ignoring it.⁩ ⁨⁨21⁩ ⁨hours⁩ ago⁩:

  Because what we call intelligence (the human kind) usually is just an emergent property of the wielding of various combinations of fist or second-hand experience by “consciousness” which itself is…

  What we like to call the tip of a huge fucking iceberg of constant lifelong internal dialogues, overlapping and integrating experiences all the way back to engrams/assemblies/memories so deep we can’t even summon them any longer but are still measurable, still there.

  Humans continuously, reflexively, recursively tell and re-tell our own stories to ourselves all day, and even at night, just to make sense of the connections we made today, how to use them tomorrow, to know how they relate to connections we made a lifetime ago, and how it fits in the larger story of us. That “context integration window” absolutely DWARFS even the deepest language model, even though our own organic “neural net” is low-power, lacks back-propagation, etc etc, and it is all done using language.

  So yes, language is not the same as intelligence (though at some point some would ask “who can tell the difference?”) HOWEVER… The semantic taxonomies, symbolic cognition, and various other mental tools that are enabled by language are absolutely, verifiably required this massive context integration to take place.

• ⁨Comment⁩ on ⁨'The Truth Is Paywalled.' Internet Vets Lament the State of the 'Open' Web⁩ ⁨⁨3⁩ ⁨weeks⁩ ago⁩:

  Agree. To take some burden off contributors, maybe we could automate some of that?

  Most of us have seen bots used for routine post processing like:

  • converting AMP links
  • finding/generating archive pages
  • exposing original AP/Reuters source
  • adding DOI source for pop sci
  • alt frontend links
  • content-aware wiki refs and the like

  We wouldn’t necessarily need traditional bot comments since our software is open. Content helpers could run during post creation, for example. My point is just that there’s existing logic for this kind of stuff.

• ⁨Comment⁩ on ⁨'There was almost a utopian feeling to it': How StumbleUpon pioneered the way we use the internet⁩ ⁨⁨11⁩ ⁨months⁩ ago⁩:

  The web of 2001 was a web of sites … Now it’s a web of platforms.

  This is all true and I agree, but a big piece of the story that’s missing is access. People online were mostly demographically adjacent. The walled gardens sprang up for the same reasons you see more walls in places with high economic disparity.

  The “utopian” attitude described in the article was a natural response to lack of representation. Society wasn’t actually as far along as millennials thought.

  That utopian vision is achievable. There’s just more work to be done than we thought.

• ⁨Comment⁩ on ⁨Facebook, Instagram, and Threads are down⁩ ⁨⁨11⁩ ⁨months⁩ ago⁩:

  Their LLMs must be cybering

• ⁨Comment⁩ on ⁨Data brokers may be banned from selling your social security number⁩ ⁨⁨11⁩ ⁨months⁩ ago⁩:

  As an ID number it has almost nothing to recommended it, especially compared to modern equivalents in Europe. It wasn’t meant to be a national ID number. The only reason we use it for identification in the US is that Americans are phobic to mandatory census.

  Obviously they’re fine handling all that information to data brokers, however.