Yes, definitely not perfect (and many other non-human attmepting to be human flaws). But the manual approach described I was replying to won’t be any better. The engine is the same.
Kind of good news (well, at least regarding the bias problem): Elon Musk says he is working on his own AI chatbot called 'TruthGPT' | ZDNET
I think this one’s a bit too optimistic, though: “An AI that cares about understanding the universe is unlikely to annihilate humans.” 
Well, maybe not intentionally. Does anyone going to the forest to study wildlife deliberately step on insects? It just happens.
And what if the AI likes the universe more than humans? What if it concludes that less humans are better for the universe?
2 Likes
Let’s dispense with all of this and call it SkyNet, because that’s what all this is, right? One piece of data at a time. Control the flow of spice and you control the universe.
2 Likes
I wouldn’t buy a car from Musk, I’m definitely not going to trust him to get AI right.
(Did you see the news recently about Tesla employees sharing cockpit video around?)
3 Likes
I read about it recently. Weird. Tesla stopped receiving such footage (without consent) some years ago. Don’t know if filming Elon’s garage sped up this decision-making process.
The crazy thing is… That’s one of the few scenarios where replacing humans with AI would be beneficial: Combing through a lot of data without poking fun at users.
1 Like
As a side note, a prominent tech YouTuber was recently hacked and had all his sites taken down and replaced by Elon Musk in online conferences. Maybe young Elon was to blame, testing the waters for a YouTube buyout.
1 Like
Although I do not have any hard data, I’m going to make a stab at what I hope is an educated guess at where the power of Almost Intelligence is most prevalent.
AI, as it is called, must not be underestimated, but after what I have seen, also should not be overestimated in the present state. There may be areas of excellence for AI that I have not seen and are not aware of, as I lack the power of millions of computers to search for and find answers at the same level as AI. Given the caveats, here are my thoughts (and they are just thoughts) of a distinctly finite human.
I see two areas of incredible power with AI:
A) The ability to test a wide range of scenarios toward a defined solution, e.g.: “‘The entire protein universe’: AI predicts shape of nearly every known protein - DeepMind’s AlphaFold tool has determined the structures of around 200 million proteins.” ‘The entire protein universe’: AI predicts shape of nearly every known protein
The protein challenge was perfect for AI. There are guidelines and rules that proteins need to follow when forming a 3-D structure. Even with those rules, however, the number of possible structures that can conform to the actual structure is, for humans, an extremely large number.
Enter AI with the power of all those computers to crunch what for humans might appear to be an unimaginable number of possible permutations for those structures.
In summary, it would appear that AI is most powerful where there exists:
A clear and precise objectiveClear rules to proceed toward that objectiveUnlimited computing power to crunch scenarios toward that objective
B) Next we have the language model, which can to a lesser extent follow the same strategy as outlined in 1. and 2. above, except that AI appears to limited by some combination of a less well defined objective and rules to meet that objective.
My guess is that AI attempts to compensate for the less well defined objective and rules to meet that objective that exist for language models. Instead, AI attempts to use the brute force of enormous computing power, leading to the hallucinations that I witnessed first hand.
This is all a guess based on the extremely limited power of a single, flawed human brain, running with data points that can be counted on my fingers and toes, and not much else.
Even though I don’t have any more fingers or toes to count on, I’m hoping to keep watch on what is likely the very early days of Almost Intelligence, that possibly, someday will become full fledged AI.
Just like the initial entrepreneurs who fantasized about the potential for power and energy that they dreamed was possible with fossilized carbon, they could not have had any idea where that power and energy would lead, including the ubiquity of:
- cars, trucks and the roads and highways to connect entire countries and continents
- extensive electricity generation and use
- ubiquitous steel and concrete manufacture and utilization
- aviation across the planet
- space travel to the moon and beyond
- a revolution in the production of food
- two world wars and regional conflicts too numerous to count, causing
- death and destruction beyond anyone’s pre-20th century imagination
coupled with
- the ability to live one’s life beyond anyone’s expectation or imagination, and
- an emerging existential threat for humanity
It took a century-and-a-half for all of the above to come to fruition. I suspect AI may offer similar opportunities and threats at a full order of magnitude faster rate than what we experienced with fossilized carbon.
Buckle Up!
scrive
I don’t know who this Al fellow is, but he’s getting a lot of work at the moment as I see him being quoted all over the place.
1 Like
Even so far as, like any tabloid hack even if they work for The Times or Telegraph, to making stuff up about Michael Schumacher.
If you’ll be my bodyguard
I can be your long lost pal
I can call you Betty
And Betty, when you call me, you can call me AI
1 Like
@hovardmiles With whom and on what are you agreeing as there are some very contrary opinions posted here. If it is with the OP then he won’t see it as he’s long gone off in a huff (hence his anon* id).
He was made famous in a song by Paul Simon back in the 80s called, “You Can Call Me Al”.
Yeah, so glad to have rushed through that era.
1 Like
My limited opinion would be, why would you not want to be conscious of what is good for human kinds but I guess Woke is in the eye of a ‘bad’ character like you explain very important and I agree with you. Interesting points you make indeed
See this
1 Like
In a quick scan through this LONG thread, I couldn’t see this mentioned. Amazon now ask if AI was used in the creation of a work when you submit your book to KDP.
They also supposedly are scanning works looking for evidence of AI being used and not declared. At this point it doesn’t appear they will ban a book using AI, however, what their next step may be is anyone’s guess.
Would you prefer such books to be banned?
Completely written by AI, yes. It’s not the Author’s work.
Ok, what percentage must the author write himself 30%, 50%, 80%? I’m just interested in your opinion. 
Mm, you have made me think about my work. I don’t write books. I just write texts. 90% of the information in these texts comes from other people who tell me. I could never have found out this information myself, because it is only in the heads of these people. In a way, these people are my AI. In this case, is the work (the result) mine?
That’s important. Facts are facts (well… but let’s not open that can of worms here). Water is wet, no matter if you experienced it, read about it, someone told you, etc.
What about the other 10%? “Just” more information you know? In the end: Still 100% information. Doesn’t really matter if a human or a machine compiled them. From a reader’s (my) perspective.
Are those 10% your conclusions, your spin on it, opinions, feelings? That’s where it starts to get interesting…
But fiction is a totally different ball game. If I can learn something new through fiction, great, but ultimately it’s not about the world, rather the human experience of the world.
I’m not interested in reading a machine parotting humans. I don’t want any books banned, but I want to know what to avoid.
I used to hunt for computer wallpapers once in a while. Now most new ones appear to be “A.I.” generated. Sometimes looking impressive at first glance. But ultimately it’s all the same nightmarish hollow and soulless crap.
Some day there will be something like a “machine’s experience of the world” and that could turn out to be genuinely interesting, too. Some day.
In my case, it is just information. A claim: “It will rain tomorrow”. You only find out later whether it is true. Most of the time it is, otherwise … Nobody is willing to pay for information that later turns out to be wrong.
Yes, or I evaluate the information of others and explain to my client its relevance and the likelihood of something happening. Facts are facts, water is wet. That is true. But what does that mean in a specific case? Does someone not jump into the water because it is not in his interest to get wet? Or does he jump into the water precisely because nobody expects him to? Facts usually take on their true meaning in the right context.
A very interesting thought. Let’s say I’m writing a story about car racers, but I don’t even have a driver’s license. Am I then the machine that parodies the racing driver? Or is it enough that I am a human being, even if I have no idea about the subject and have only learned everything from machines (Internet)?
Also interesting. What happens when machines talk to each other about people, just as we now talk about machines? 