It’s satisfying to see big companies scramble to avoid being disrupted in a market they have dominated for so long…and doing it so wrong. I care about accuracy in AI, but the kind of “safety” they’re concerned about—the AI not hurting peoples’ feeeewwwwings—is definitely holding them back.

“In Paris demo, Google scrambles to counter ChatGPT but ends up embarrassing itself”

https://arstechnica.com/information...

Also, people should stop viewing those AI assistants as something that spews out facts.
It’s about assisting you in finding what you are looking for. Not judging what’s true or false.
The wrong expectations are holding it back.

@Lensflare Maybe that isn't a bad thing right ?

I mean, ask ChatGPT if the earth is flat, it will state that it isn't. So a lot of people may accept from some advanced AI that this is a fact.

Problem will be when those facts are altered based on the vision the creator has.

Maybe one will be more religious and will alter the output of the AI to favor that belief.

This is in some kind, what happens to schools. We all learn stuff at school from 'teachers'. I did go to school in France and in Belgium and trust me, some facts about some historical stuff is twisted in favor of the country you are living in.

Stuff like that is happening all over the world.

@Grumm Like Egyptology. You have some people claiming their ancestors built the pyramids and even believing they are superior genetically as a result. But when you look at the hieroglyphs you find very light skinned people directing dark skinned people in some of the reliefs. And in one case a blue skinned person. They have also done some genetic testing and found some of the later pharaohs were of European descent. I think the people looking with a less biased eye are finding Egypt was very cosmopolitan and changed over time. Another fucked up thing about this is a lady tested the pharaohs and other nobles with drug tests. Found they had traces of cocaine. Cocaine is a South American plant. How the hell did they get that? There are a ton of researchers that have contradicted the official Egyptologist position, but everybody has a view that won't change. The worst is the people throwing racist elitism into the mix.

@Grumm Putting effort into improving the assistant AI in terms of facts and truth is the mistake, imo.
I know that the companies try to be vague on this because there is competition, but this false advertisement or false impression of the users make the product worse.

Even if it would be technically sensible, it still would be immoral. AI should not replace teachers. It should’t be teaching at all.

@Grumm Here is the exact opposite. Everything is being skewed away from any sort of cultural support for anyone who believes there is a God or that Western civilization is a good in the world. The AI experiences I’ve had include disparagement of anything traditional and outright religious-style indoctrination of everything far left. Dozens of queries I’ve made about contentious issues end up with answers falling far left of center on cultural, political, economic, and religious facets.

I wouldn’t at all be surprised to find out that OpenAI is an intelligence community front for swinging individual cultures toward desired ideological outcomes so that it is easier to govern us all globally with zero resistance.

@Demolishun Spot on. You’re my kind of people! Zahi Hawass and his cohort have made an absolute mess of Egyptology. And archaeology globally.

@stackodev What blows me away is the blatant denying of what is in front of them. The Sphinx looks like it was carved to look like an African descent (black) person. But they say it looks like some pharaoh and all the dimensions are wrong.

@Demolishun And it’s WAY older than they state in all the museums and literature. Look at the erosion at the base. That didn’t happen with wind, but with lots and lots of water flowing by it. That means that the area was subject to multiple floods. That kind of water didn’t flow there in 4,000 BC and later. It had to have occurred when the region was greener and that would’ve been during the Younger-Dryas period and before.

@stackodev Yeah, like 15k to 20k old. Crazy.

@stackodev For me, most of all the events are not correct.

They found viking ships in canada... So they were the first to discover america. Why isn't that not a fact today ? Are we all afraid to change the history books ? Just because we have computers, doesn't mean any previous civilization was inferior is now.

Ancient Egypt has so many secrets. Same as Maya's or any lost people who made great buildings.

@Grumm There’s so much that has been buried or withheld. I have made a study of it for at least 25 years. I could talk for hours about it. The Mayan people have a book that survived the codex-burnings which they had been transmitting orally for centuries. Two Mayan priests dictated it to a Catholic monk in the 1500s. They directly state that at least a part of their people came from the Middle East on boats. Archeologists and anthropologists just throw that part away because, well, why believe “the natives” about their take on their own history? Better to correct them and retell it for them in correct form.

It’s not a hard guess. Generally, companies don’t want to have the next Microsoft Tay, that became a hate fueled machine after 24 hours in Twitter. So they attempt to put locks on their stuff.
I bet you’d hardly mind if whatever AI you’re complaining about was spitting out things you like, regardless of whether it was truth or not.

@Grumm I watched a documentary called "The First Americans". It was aboriginals many years before anyone else. Like 10s of thousands. They found cave drawings showing them fighting invaders when the Asians came over. The Asians became the Native Americans. They found 2 surviving women at the very tip of South America that were aboriginal descent. It was a really fascinating show.

They also found Chinese dynasty pottery in Inuit tribes in Canada. How they got that is probably another fascinating story.

@ars1 lol, yes, this would happen just talking to employees I think. I would definitely give the AI "some ideas".

@ars1 Thanks for putting me in a box I don’t belong in. I love the free exchange of ideas. I have debates with Voltaire-style liberals and enjoy them. It’s the authoritarian Leftists that seem to control every thought and ideology today that I don’t want running whatever AI turns into.

@stackodev You say that, but immediately skipped over the main point of my comment. If you want an unfiltered AI you can have a go at it yourself. What could possibly go wrong?

@ars1 Isn't that what Copilot did ?

It looked at all the public code on github to make his own answers. All fine until thousands of people created false public repo's and the AI started using those as 'correct'

Doesn't that reflect the real world ? If enough people believe in one thing, (true or not) it will become true anyways.