ChatGPT: how Google dies

[Jeff]

You will live how they say

Nope

You will live where they say

Wrong

You will live when they say.

Wrong again

You will get the healthcare and vaccinations/procedures they say.

Definitely not

You will only travel where they allow.

Hell naw

You will eat the foods they give you.

absolutely not.

[Dan_AKA_ROY23]

ChatGPT is surprisingly susceptible to error: Thompson pointed out how when asked if philosopher Thomas Hobbes believed in the separation of powers, the supposedly omnipotent AI incorrectly invoked an argument by Hobbes' contemporary John Locke.

It also frequently flubs three-part mathematical equations, such as claiming that 4839 + 3948 – 45 = 8,787. Thompson attributes this glitch to the fact that it's programmed to match data patterns rather than calculate numbers.

According to the Times of London, "its knowledge base only goes up to last year." And OpenAI admitted the bot can give "plausible-sounding but incorrect or nonsensical answers."

[Dan_AKA_ROY23]

Can it translate languages accurately? Google Translate (DeepL and others) notoriously struggle with idiomatic expressions. The barrier between different languages may never be solved.

[Dan_AKA_ROY23]

Once you start splitting language into regional dialect....and then throw in accents....nothing has a chance!

In Spain they have a weird lisp sound for c's and z's (sounds like our 'th' sound). Hard c's are the same for Spain and Latin American countries. The Spaniards use 'vosotros' (2nd person plural you). Latinos NEVER say it. Various expressions are different depending on which Spanish country. The "ll" sounds are different (Spain and majority of L.A. countries have a "y" sound, Mexico uses a "j" sound, Argentina uses a "sh" sound). And Costa Rico uses "vos" (2nd person singular you).

One of the issues learning Spanish...it varies so much.

[mowens]

Bring on the robot overlords

[TylerBeer]

According to the Times of London, "its knowledge base only goes up to last year." And OpenAI admitted the bot can give "plausible-sounding but incorrect or nonsensical answers."

ah! so they invented a republican

[mkd]

[mkd]

Just signed up for an account. Extremely interesting.
 Using it to debug my Google sheet.
 Pretty amazing

[mkd]

ah! so they invented a republican

.lol, as opposed to a demonrat that is lacking the plausible part. 🤣

[Here's Johnny!]

For the doubters remember it is constantly learning... it will make mistakes at first but if you correct it it keeps building intelligence.

Crazy thing is for students...easier ways to cheat!

[Here's Johnny!]

This is a research group using AI to find new battery materials.

https://www.theguardian.com/whats-next-ask-toronto/2022/oct/25/vitamin-powered-batteries-how-self-driving-labs-are-using-ai-to-develop-remarkable-new-materials

[kccadcam]

I've been tinkering with Home Assistant, running on a Raspberry Pi with an SSD.
I belong to the Home assistant group on FB, this popped up from one of the guys messing around with ChatGPT to write scripts:
"So ChatGPT offered to integrate with my Home Assistant instance and automatically
create automations based on machine learning and predictive algorithms. Anyone else try that yet?"

Not a bit scary at all.....  

[pmartin]

I think it takes a lot of credibility away from artificial intelligence when u see things like self driving cars running people over and crashing, the technology just isn't there yet..

Running down the enemy or committing suicide to escape subjugation actually sounds like something that a human might do.

[CNCAppsJames]

I asked it simple easily verifiable though any search engine, random 80's trivia questions and it got the majority of them wrong.

The kicker though, there were elements of the answer that were correct. The scary thing is people that don't know any better or people like Smithead that are too fuckin lazy to do their own research will say "but ChatGP said". Thus a new truth is born and the old true and verifiable historical facts of record will die.

AI is how civilization ends.

[Incogneeto]

I asked it simple easily verifiable though any search engine, random 80's trivia questions and it got the majority of them wrong.

The kicker though, there were elements of the answer that were correct. The scary thing is people that don't know any better or people like Smithead that are too fuckin lazy to do their own research will say "but ChatGP said". Thus a new truth is born and the old true and verifiable historical facts of record will die.

AI is how civilization ends.

I don't know about that.


Maybe we can tweak it. And thereby Tweak Smit.

An AI that can't Lie.?