101010.pl is one of the many independent Mastodon servers you can use to participate in the fediverse.
101010.pl czyli najstarszy polski serwer Mastodon. Posiadamy wpisy do 2048 znaków.

Server stats:

488
active users

#fab

0 posts0 participants0 posts today

Am utterly entranced by my first tech device I am literally playing with. There are features I didn’t expect that are quite pleasant. Got a new Apple iPad 11 on the Memorial Day sale—$299 +Apple Care 2 years $368. It had been years, maybe 9, since I touched an iPad as part of my job then. The works inside this one are like an iPhone 15. Didn’t expect the camera and magnifier to be so good. Didn’t know the graphics scale up great from what I am used to on an iPhone. And yes, I am playing Clockmaker on the new big screen with a new appreciation of the artistry in it. Photo is using the quick magnifier view on my desktop system, foreground shows mouse mat/playmat with the Michael Whelan art. Keypad and set of 4 fave Flesh and Blood cards, Last Ditch Effort. That magnifier has a built in camera—very nice. @MichaelWhelan #ipad #Whelan #FAB #Fantasy #FantasyArt

Replied in thread

#AI are #FAB (Free Association Bots), and are nothing fabulous.
@OliverUv wrote:

The inability to learn from small amounts of data is a fundamental flaw...

Broadly, yes. It takes a human child a decade of experience building and sorting memories for value to begin to "simulate" a human adult. Remember how interminable life was back then? Everything was new. That's a human in "learning mode."

... that sets modern "ai" apart from human intelligence

Nope. The difference is that child is also learning to reason. The current flavor of AI is not.

While LLMs are expensive to train both from an energy standpoint and an ethical standpoint, the fundamental flaw is not that they cannot learn from small amounts of data (well, once it is built, like a human child is "built.") It's that they can't reason enough to reject data, analogously to how a child can be indoctrinated into hateful or anti-social ideas.

The concept of AI poisoning utilizes this flaw. You and I can demonstrate the flaw "benignly" by observing how often-updated AI give trending answers. Not the best answers but the trending answers.

An AI is a programming component that returns information by pattern matching and frequency.

Here's where the bubble of uselessness is: No reasoning ability.

Programmers add reasoning when they use AI for targeted applications, and then they have to curate the results.

With no reasoning ability, a generative AI is equivalent to a person answering questions by free associating and saying anything that comes to mind, no matter how absurd or unreasonable or uncensored.

A "free association bot." Let's coin the acronym FAB.

That people believe an AI can do things for them without knowing how to do that thing in order to verify the returned answers are reasonable—that a FAB can write for them or research for them or replace employees that they have to pay to do work for them—the whole thing is going to collapse the same way that it would collapse giving a ten year old with a good memory the job.

cc: @freeagent @vextaur

#chatgpt#llm#llms

Die meinen tatsächlich, dass Mobbing eine Form von politischer Kultur sei.
Dass die Bevölkerung gleich reagiert, wie sonst auch, wenn man jemanden "fertigmacht", ist dem Herdentrieb zu verdanken. Der Mensch hat gerne die gleiche Meinung, wie die Anderen, je mehr das gleiche tun, desto einfacher wird es mitzutun.
Es ist einfach so, dass entgegen der landläufigen Meinung, immer Menschen gefunden werden, die ohne Zwang Henker sein wollen und es mit Freude sind. #FAB fight against billionaires

Interessant ist, dass in Schweizer Medien über die Anzeige von Habeck gegen einen rechten Hetzer berichtet wird (unter Zitaten der AFD) aber nicht über die "Aktion" der FDP. Wer sich jetzt fragt warum das so ist, müsste nachsehen welche Aktionäre und Geldgeber hinter den Medien stecken.
Der neue Slogan #FAB (fight against Billionaires)

Skończyła się faza swiss Mistrzostw Świata w i wyłoniła się czołowa ósemka, która powalczy jutro rano (rzecz się dzieje w Osace) o tytuł Mistrza Świata.

Mamy Polaka w tym gronie - jako ósmy zakwalifikował się Grzegorz "Urlich" Kowalski, gracz znany w przeszłości z MTG, a który tutaj gra Enigmą :)

fabtcg.com/en/coverage/world-c

fabtcg.comFlesh and Blood TCG