Home | Notifications | New Note | Local | Federated | Search | Logout
Federated Timeline
Reply to @lain@lain.com
Blurry Moon@sun@shitposter.world (2026-09-23 07:05:42)
@lain the first zoomer
Reply to @jeremiah@shitposter.world
Blurry Moon@sun@shitposter.world (2026-09-23 07:05:20)
@jeremiah @Omega_Variant @cjd @meowski @nimrod I odn't think you have to to prove it learns. at amy rate II think our disagreement is I'm not bothered if it doesn't work exactly like a biological brain if you can prove it practices reinforcement training for example. the irony here is, this isn't how LLMs work at all, and if you look closer at "reasoning" LLMs, they do in fact do something that isn't really like human reasoning at all and arguably isn't any real kind of reasoning at all either.
lain@lain@lain.com (2026-09-23 07:04:18)
Thinking about listening to this but I’d prefer not to
---Attachments---
image: https://wired.lain.com/media/5e/d2/e4/5ed2e4796633106b04d4da9278ee6a54efb0d18e5af3571c78b88c8e435b2329.jpg
Reply to @sun@shitposter.world
Jezza™@jeremiah@shitposter.world (2026-09-23 07:02:27)
@sun @nimrod @Omega_Variant @cjd @meowski but how would you prove the artificial learning process is equivalent to the fly's? By behavior alone? How would you test the simulated environment against real flies?
classic mametschko@mametsuko@mk.absturztau.be boosted:
@samy@eupolicy.social (2026-09-23 05:18:54)
Feeding time is a serious business!
#catsofmastodon
---Attachments---
image: https://eupolicy.social/system/media_attachments/files/117/316/539/583/962/152/original/aa791d26ec01e3dc.jpeg
Reply to @jeremiah@shitposter.world
Blurry Moon@sun@shitposter.world (2026-09-23 06:59:17)
@jeremiah @Omega_Variant @cjd @meowski @nimrod they are not substitutable because one is an extremely pared down and adapted to binary math. yes the brain has layers of stuff that aren't neural nets. you are not going to fully simulate a complex brain in perceptrons for many reasons.
classic mametschko@mametsuko@mk.absturztau.be boosted:
@mametsuko@mk.absturztau.be (2026-09-22 17:49:02)
today's gacha :maomao_comfy:
---Attachments---
image: https://misskey-taube.s3.eu-central-1.wasabisys.com/files/webpublic-5f022517-021d-4e4b-995d-824c75d1182c.webp
image: https://misskey-taube.s3.eu-central-1.wasabisys.com/files/webpublic-6c43003a-8dc8-4ec9-b040-a3008cb7a5ca.webp
Reply to @sun@shitposter.world
Jezza™@jeremiah@shitposter.world (2026-09-23 06:56:23)
@sun @Omega_Variant @cjd @meowski @nimrod there's more to it than that. Brains are not just and only neural neworks. There are other systems at work including non-neural systems that effect the functioning of the brain.
Digital neuron to brain neuron is a very weak analogy, with a only a conceptual similarity. They do not function equivalently in that one may be substituted for another, or that one could do the other's function.
Reply to @nimrod@shitposter.world
Blurry Moon@sun@shitposter.world (2026-09-23 06:55:42)
@nimrod @Omega_Variant @cjd @jeremiah @meowski there are entire classes of drugs where we know they work but we don't know why they work. we can feed the fly brain into a computer and simualte an environment around iit and it navigates it and learns
Reply to @jeremiah@shitposter.world
Blurry Moon@sun@shitposter.world (2026-09-23 06:54:43)
@jeremiah @meowski @Omega_Variant @cjd @nimrod perceptrons are not neurons but you are splitting hairs that it's different so it doesn't learn, when it was designed to maintain the quality of learning, which it does.
Reply to @sun@shitposter.world
nimrod@nimrod@shitposter.world (2026-09-23 06:54:38)
@sun @jeremiah @Omega_Variant @cjd @meowski I know we have the map of the entire fruit fly brain but do we actually know what every component of it does? Do we know every biological function it controls and how it interprets data from the eyes and ears?
Reply to @meowski@fluf.club
Jezza™@jeremiah@shitposter.world (2026-09-23 06:53:33)
@meowski @Omega_Variant @cjd @nimrod @sun brains are not neural networks in the same way an ANN is. This is another semantic hinging case where the terms "neural" and "neuron" were chosen, even though there's only a basic conceptual similarity between organic brain neural systems and software neural networks. A digital neuron != a brain neuron, and an ANN != the brain.
From a pragmatic perspective, sure, these tools can do jobs that previous software systems could not do. Somewhat. With a great deal of supervision, overhead, catastrophic and comedic misfires, and eventually deskilling people.
Reply to @gabriel@mstdn.starnix.network
lain@lain@lain.com (2026-09-23 06:52:18)
@gabriel this post is made up 100%
Reply to @jeremiah@shitposter.world
Blurry Moon@sun@shitposter.world (2026-09-23 06:51:47)
@jeremiah @meowski @Omega_Variant @cjd @nimrod computer neural nets are a mathematical approximation of a component of biological brains and the definitely do learn. they don't learn as well as a complex mammal brain because a mammal brain is a network of many neural nets dedicated to specific purposes. entire fruit fly brain has been simulated.
Reply to @jeremiah@shitposter.world
woodland creature@meowski@fluf.club (2026-09-23 06:49:55)
@jeremiah @Omega_Variant @cjd @nimrod @sun ok i didn't say AI was a brain, but brains *are* neural networks. it's more than semantics. also look at the definition of reason: "think, understand, and form judgements logically" the understand part is what's really in dispute. i would argue that understanding concepts is a requirement to engage in reasoning, and they demonstrate that ability.
forget about marketing. i'm looking at this purely from the perspective of what they can do. what useful work can they perform, based on my own experience.
Gabriel@gabriel@mstdn.starnix.network (2026-09-23 06:46:33)
I'm assured that another round of funding fixes this. Just one more. Maybe two to be safe. Three rounds should be more than enough.
---Attachments---
image: https://s3.us-west-002.backblazeb2.com/mstdn-starnix-media/media_attachments/files/117/316/880/857/388/925/original/1ec53fe10af2f40b.jpg
Reply to @meowski@fluf.club
Jezza™@jeremiah@shitposter.world (2026-09-23 06:46:12)
@meowski @Omega_Variant @cjd @nimrod @sun in my opinion, you are making a semantic argument, hinging on the fact that words like "reason" and "intelligence" are appearing in two very distinct domains. The confusion between those domains is mainly a product of marketing.
Neural nets are not equivalent to brains, statistical "learning" on neural nets is not equivalent to reasoning, and the whole of it is not intelligence.
some cat@somecat@toot.cat (2026-09-23 06:45:38)
[toot]
#tootsday #toot #fediArt #art #blobcat #blobfox
---Attachments---
image: https://pool.jortage.com/tootcat/media_attachments/files/117/316/869/439/733/047/original/5707d7eab0953899.png
Reply to @jeremiah@shitposter.world
woodland creature@meowski@fluf.club (2026-09-23 06:42:00)
@jeremiah @Omega_Variant @cjd @nimrod @sun they absolutely do have reasoning abilities. computers excel at logic. they literally contain billions of logic circuits. abstract reasoning is an extension of that.
i will gladly look up the definition of intelligence for you: the simplest definition is "the ability to learn"
neural networks absolutely do learn, they can be "trained" with reinforcement learning.
Reply to @xirin2u@mastodon.social
Urei@LLENN08@hiradai.space (2026-09-23 06:41:45)
@xirin2u@mastodon.social seems fun fun dream
Reply to @meowski@fluf.club
Jezza™@jeremiah@shitposter.world (2026-09-23 06:36:55)
@meowski @Omega_Variant @cjd @nimrod @sun they don't have "reasoning" capabilities unless you stretch "reasoning", the same way they do not have "intelligence" unless you stretch the definition of "intelligence."
Perhaps start by defining intelligence, so we can distinguish between it and artificial intelligence.
evacide@evacide@hachyderm.io (2026-09-23 06:34:40)
German agencies use the linked-device features in WhatsApp, Signal, and Telegram to receive messages without breaking encryption. https://cybernews.com/privacy/police-telegram-whatsapp-signal-surveillance-linked-devices/
Reply to @jeremiah@shitposter.world
woodland creature@meowski@fluf.club (2026-09-23 06:33:56)
@jeremiah @Omega_Variant @cjd @nimrod @sun this idea has existed for a long time and now it's real. what else are you going to call man-made neural networks that have natural language conversations and reasoning abilities? that's exactly what intelligence is. machine learning = machine intelligence.
あゆみみみ:internet_oekaki_man:@are_you_333@misskey.io (2026-09-23 06:29:11)
リクもらったキャラのラフ描いてみてるけどムズカシイ……
お姉さんキャラは苦手だ:blobcat_frustration:
---Attachments---
image: https://media.misskeyusercontent.com/io/a6165c3d-8a78-435c-bf33-3dfc98377a96.webp?sensitive=true
Reply to @lain@lain.com
Blurry Moon@sun@shitposter.world (2026-09-23 06:25:57)
@lain @kaia if you're writing it in an expressive language you don't have to be constrained by a shitty c design
Blurry Moon@sun@shitposter.world (2026-09-23 06:25:07)
@hfaust I just can't get mad about this. get peppered, fuckface
Reply to @feld@friedcheese.us
Blurry Moon@sun@shitposter.world (2026-09-23 06:24:19)
@feld "when things are bad"
Only 3 Easy Payments of $19.95@feld@friedcheese.us (2026-09-23 06:20:01)
LLMs can't (yet) invent new words like "enshittification", so Cory Doctrow's job is safe 🎯
Techokami@techokami@woof.tech (2026-09-23 06:19:35)
People on fedi: We need a new C compiler to get away from what's happening to GCC!
Me: what about improving tack
People on fedi: what
https://tack.sourceforge.net/
The only issue is that it doesn't do x86-64 but, ehhh, it already targets the important stuff like SPARC and the 68K family and PowerPC and i386 if you really need x86
Reply to @kaia@pleroma.soykaf.com
lain@lain@lain.com (2026-09-23 06:18:52)
@kaia https://github.com/asterinas/asterinas blink and you'll miss it
Older Notes