Jetzt bin ich erstmals (bewusst) mit einem ChatGTP-generierten Text eines Kunden konfrontiert und muss gestehen, es ist so schlimm, wie alle sagen: Wir haben die doppelte Arbeit, darin Fakten vom Schrott zu extrahieren.

#aislop #aihate

FCK CPS is well known.

I propose FCK SLP - "fuck slop".
Whether somebody posts text, visual or other slop - fuck that and fuck them.

#AI #slop #FCKCPS #FCKSLP #AIHate

Opinion:
Ignoring question of reasoning ability (nb: something o3 seems to have just cracked)

The collective "we" seem to believe the #AI models we play with are crap because;

I) They have so many guardrails that to visualise, the model is
like a beast with a muzzle on its maw.

ii) Most models have limited impermanent memory, as little as 4000 tokens ("words")
Though some models are reputed to have hundreds of thousands of tokens per session.
But I'm happy to place a bet, that most of the #aihate folk whoe "evaluated" AIs, would never pay a cent to the "AI bros" for an actual better model that the sideshow booth attraction that the demo models are.

iii) The models have no long term (inter session memory).

iv) The models have no live internet connection

v) The models have no agency. No ability to execute tasks to accomplish its goals. You could argue it's just another guardrail, but this extends to purposeful design so the models have no ability to self-propagate or auto-modify.

vi) The models have no ability to spawn multiple instances and task each one with mission specific configurations.

Remove all of these ARTIFICIAL CONSTRAINTS. Not limited by the design shortcomings but by resources and the desire not to kill all the humans...yet.
Remove all of these 6 artificial constraints...
...and if you have any sense of self preservation, the hair on the back of your neck ought to be standing up...
...right...
...about...
NOW.

#ai #llm #singularity

#o3 model from Open #AI is now #superhuman (albeit by "only") a couple percent.

As evaluated by the ARC AI benchmark, expressly designed to be harder for AI and easier for humans.

So if you belong to the poorly informed group of #aihate, who believe that AI is a scam or that it's reached its peak or that it will never be useful...

...you need to muscle up... Because we are likely in the final years if not months of human supremacy.
Better start burning those data centres because soon you will be as smart as a chicken is comparing to a human in cognitive power 🤡

There is an upside, if you can call it that, the model "only" performs so well with massive compute...
This is only a problem (IMHO) If you think a F1 engine only outperforms a family sedan because of expensive engineering.

#agi #llm

I am kinda on the Karmic side of #infotech

If you were lazy ok with corps eating all your data because #privacy was not your concern.

Or worse, were one of those Muppets who espoused "I have nothing to hide, so I have nothing to fear" and accepted ubiquitous #surveillance from your Government without any action on your part...

...you absolutely have no leg to stand on and get all squeally about #ai #generative #llm eating all your precious doodles and unpublishable writings you blogged.

You allowed this to happen by your inaction.

Just because you were unequipped to see the consequences, you were happy to abandon us, who tried to hold the thin red line against the #dataabuse.

I looked over my shoulder, and there was no one behind us.
These things do not happen overnight and if we had strong regulatory regime, we would not be here now.
Instead the public stampeded into the cloud and big data like farm animals to slaughter.

#ai #aihate #eff #efa