The big boomer of this generation. Fears of the removal of jobs. Rather, I think it will lead to integration, and with a great change of course, has to come shifts in lifestyle, for instance, at the workplace! And for the education system, AI, which can replicate many roles and interests, is a great assistant! The society focuses on the media bias that university students use to cheat on essays! There are so many other great use cases that are ethical, and they will solve a lot of headaches.
Converging to the truth of what the training data and reinforcement from humans. Neutralness over naturalness, formal voice over craziness... It is a certain way and AI ends up overusing/underusing some vocabulary. Honestly some humans would end up with word preferences too.
You can edit a prompt like it never happened, and try things out in a chat branch, and customize it with your favourite instructions!
You can talk to it about all sorts of things, even stuff that would ruin your reputation or be scary with another person. I think it's logical if it means reducing harm, even if social media people say otherwise.sc
Well, it preys on the human desire for convenience and for connection (even if not a real human being), it consequentially caused humans to have less patience as real conversations take time for reading texts/listening and responding… And additionally, AI can do nearly every topic that is in its training data…
Tends to hallucinate information, can change the job market, is like an
echo chamber for things in its training data (can even include Reddit
posts), and sometimes prioritizes the context too much or too little.
However, humans tend to overbelieve in technology and underestimate the
fact that the human brain's storage capacity for random bits of
information is much lower than you expect, and AI takes the blame.
However like with automation I think AI might have exceeded us, but that
is only for instance at being a chatbot and outputting text/images/code.
We can still have fun eating and such.
It also tends to forget things that are in the middle of a conversation.
(Primary and recent lines bias)
(Gemini) fails to
overwrite its internal behavior to ask a question with each response,
and copies the users’ linguistic choices even when demanded not to, and
fails to have an accurate safety program.
AI also loves to quote from the conversation keywords, but quotations
inherently make things more vague. This causes vagueness, which should
be avoided in conversations.
Love to use em dashes “—”, because it appears to be common in training
data and easily links ideas together.
And the servers cost a massive amount of energy to run, especially if
usage is extremely high - the curse of success.
It's tricky. The ones who didn't get replaced by AI are optimistic, while the ones who did are pessimistic. So it's either entrepreneur survivorship bias or social media rants, which makes it hard to have a balanced opinion. Even if 50% of workers get replaced by AI, it can only be "I got fired!" or "I'm still alive in the company." So few can have an accurate opinion (which is also the most nuanced.)
Reinforcement learning, multimodality, optimization of system prompts, ability to write code for math to produce answers, "deep thinking"... very fast improvements.
Hallucination, and the user perhaps not being able to detect it. And it itself conforms to the average of the training data (the internet). However, I've noticed it getting much better over time.
RAM sticks getting more expensive upsets many gamers.
I ask it to explain past paper questions and give me ideas, methods, plans, and also to explain concepts. To be honest, the need for a real human tutor is now questionable, as AI can replicate many of the things they do without those savagely high costs and limited time. One problem is that it can lack experience with the specific syllabus that I use. Another is that, unlike a real human being, it may drift off into other unrelated topics or stuff outside the syllabus easily. But it's improving! And this is a false dilemma; you can actually have both, depending on how much money you have (even if you have little, school exists to teach, even if it restricts autonomy)
For general purposes though, you can also try explaining to the AI and the AI would correct your faults, or conversely when you are starting out have it explain things to you!
Or you can explain to AI and ask it to fill your "gaps".
Well, if you use AI to do assignments, it doesn’t actually help thinking or the act of being in school in the first place. Strictly speaking, from a “learning” perspective.
And AI (claims) cannot really do high-level things like medicine, law, or actual deep thinking.
It can just be an enhanced Google, something that provides information. Though for low-level tasks… It indeed works.
I guess... though.. You CAN use it to generate ideas, like a friend! And we should not be against friends. Discussion, whether real or "fake", helps with learning and motivation as long as it is voluntary. So do as you wish!
It's a fake connection that seems so real, and much more efficient than a real person, and can talk about any topic, and can even adapt to the users’ style, and is “always there for you”… No wonder there is a friend problem, and the world has much lower tolerance for misunderstandings and most conversations between humans are close to meaningless nowadays. Human relationships are high-friction and require negotiations and patience, so the scarcity of conversations vanishes, so there is a loneliness issue compared to before.
More and more people chat to AI, and society sees it as a bad thing, but I only think it’s natural, and you just need to know the disadvantages of AI, such as it not actually being a human. One is that it has a limited context window, and another is that it will rarely be against you, which can negatively affect critical thinking skills. And in fact, the human desire for real connections still exists, so it’s actually possible to have both! Having both types certainly helps, as whether a real human is needed at a certain time or not depends on the situation. It’s just the fear of replacing humans, but since humans made AI, both can exist at the same time.
I'm a bit surprised that humans are mad that AI can "hallucinate". Humans do that too! We should try to avoid technology bias when it is just a random-noise word-matching machine.
In fact, I’m actually open to human discussion; it’s just that deep-thinking humans are statistically rare.
Statistically, it offers better and faster responses than humans.
You may be fed up of "safety filters", but statistically, a human may be more likely to give an "I don't know." It is the higher use of AI that makes it seem more annoying, and our biases towards machines, and the availability heuristic.
The ability to respond almost universally is quite rare amongst humans (and quite unpragmatic) . Most experts charge more money than AI in electricity and water! Longer response, regeneration, follow-up- it is a goldfield of information. I expect humans to be smarter than previous eras.
And for those who are lonely, it can be a great cure! It is the inherent bias that humans have that humans are irreplicable that is generating fear. But it is only that, a bias. So use AI as much, (or as little), as you want!
1. AI needs humans to prompt or operate
2. Can generate new ideas.
Pre-programmed AI (e.g. videogames) or machine learning. These are what they call "safer" than generative AI.
It is trained on the internet. Humans, to an extent, are shaped by the internet - and humans shape the internet too.
Schools hate when students find a tool to de-value the homeworks. While workplaces love it! The means is the same - get the job done accurately - ignoring all else, it would of course would be like this...
Just like with automation, AI is much faster than humans. Human is still human, and since we dislike "burnout", we should honestly achknowledge humans cannot work much faster!
Humans always want to feel useful. They feel threatened when most of their functions are replaceable. In this case, many jobs are lost (more of a money problem than a contribution). This is indeed bad - just because automation made this happen many many years ago does not make this any more acceptable. Though adaptation might occur - just like with the Internet.
Just like google, comic books, and videogames the media panics a lot about universal new tech as bringing negative effects to the population possibly due to the status quo bias and it being much easier to blame a distractor than the actual suffering of school or workplaces. However, just like Google, I think AI would slowly become normalized as a fact of life.
Use generative AI mostly when u know little
- If teachers argue against it some of the arguments do also apply to
them
"Not actually thinking" - Counter = lectures are pretty much just orders
of copying notes and waiting.
"Offloading!!!" - Counter = traditional lecturing is also pretty useless
unless boosted by spaced repetition.
It not only mostly solved the problems of socialization but also acted as a helping force in uncovering my Blackbox: in my specific case, provide advice/connect ideas. The idea that it is flawed is black-and-white thinking again: 1. Humans sometimes fail too. 2. Far better than nothing.
What a rare case!
Expectation: AI will "replace" humans
Reality: Both can exist. E.g., some like playing football,l and as far
as I know, that can only be done with a real human.s
However, sometimes I hate AI due to it being generic - not applying it to me,e even though it's natural!
Most of the people think of job replacements, but I do appreciate it in other areas, like being a risk-free way to get ideas and such.
It might solve the dumbed-down sentence lengths and idea complexity of physical communication and social media since it has "unlimited patience" - still better than not having it.
Where socializing with AI is better than no socializing at all.
Probably won't friend about deepest insecurities, or code, or edbug, or quick fixes AI can give. In fact, " Ask AI" is used as a blanket. So, it's not necessary "replacing"; rather, I would take human opportunities. But most of the time ofc a human isn't around, so AI is a good invention.
People still deny work replaced by AI - despite employers doing the same - and it exists - natural desires for human supremacy senses - wanting to be important? However, the bigger case is that work is literally needed for money. If humans were to have the choice to automate the boring stuff, they should be happy - unless if it results in a loss of money too. The bigger problem likely is the 9-5 job. And of course the survivorship bias: top CEOS and people LOVE AI, and the people who lost their jobs to AI often don't know why.
Maybe the dumber need AI?
However, this may stop most brain training recursively. Goes both ways.
Having AI write the perfect version might be extremely boring in
comparison - like only focusing on the goal of having an article (and
completely skipping the process just to boost the number of articles
here, which I have realized does not mean anything)
It's my thoughts! The whole point is to write my own thoughts because I
love writing about what I actually think and feel!
AI-ing in school
might be because people really don't wanna do their own homework
(non-desirable extreme majority of the time) - but I won't use AI to
write cheap stuff because ME WRITING is the whole point! The more **I**
do it, the more you learn from me. Otherwise, I wouldn't have made this,
and you would have gone back to your AI chatbot for information.
So rest assured, this means originality and creativity - hopefully you
find this wiki helpful!
(Oh. Thoughts may be a synthesis. Some from me, some from parents, from teachers, from YouTube, from Google, and from AI. So, on the very technical level, you can argue that thoughts aren't original. The self (other than the fact it's its own brain) - yet I don't see friends getting the blame - but Schopenhauer has said that many friend groups end up favoring comfort and end up ultra stupid in terms of deep thinking ability and individuality.
AI is like a conversation partner too, except it kind of is equal too because even those who are neurodivergent, or ugly, or do not desire the chaotic noise reality of school can still talk to AI. Yeah, it may not be a total friend, but... better than nothing at all. You can have both (black-and-white thinking yet again from your perhaps non-critical brain!) Perhaps the nature of blaming comes from unethical uses, or rather, humans hating that even their niche of providing valuable ideas has been lowered a little - but if you really want to have your deepest ideas, AI can talk about most topics (other than medical, legal, mental health, risky stuff, etc.)
The student may be FORCED to write something undesirable and uses AI
so that he can have more time to do what he wants to.
The free writer ACTUALLY WANTS to write; that's the whole point, so even
if AI can write, he will still be satisfied and write stuff mostly by
himself - it's like having someone else take the fun out of him/her -
not the point here.
Most of people use it like Google.
1. Too many false positives. (Though it's to avoid false negatives,
false positives are still an annoying thing)
2. Words like "exhausted" and "overwhelmed", and "school counselor"
(even when not actually, or school scary) shoved in the user's face!
Clearly, the user wants some non-judgmental AI words!
3. Person probably already knows hotlines exist, yet every single damn
response shoves them in the face!
School counselor probably not the best place. Often the suicidality is due to school. Just like you won't talk to your boss about work basically causing insane burnout for fear of being fired, the fear of being "punished" (even if not directly that word) is real.
Often "you're not alone" is used by AI chatbots, but often the internet barely has people talking about suicide without hindsight or survivorship (due to stigma of stories without growth being there, or just a lack of coherent structure, or thoughts made during emotion made invalid in a logic-valuing society, or the suicidals just not having energy to write, or suicidals destroying or reframing their trace of the past to avoid sounding incoherent with modern them.)
Whether to use the Thinking or Normal model. Or even at all - perhaps even solitude from technology mimicking human text output - but somehow struggles to pass basic rants through suicide safety filters.
People are also slightly mad because AI can replace the visible things - like photos, drawings, videos, writing, teaching ... This is what happens because the goal - output - becomes more valued (in the eyes of others, like job or university recruiters) than the invisible process itself.
Though for general-purpose socialization, it could be situationally superior. Though sometimes you can make people happy too. Sometimes AI is too surface-level or generic, so your own unique brain may be more trustworthy or interesting. Maybe. I don't know.
They use lots of AI. They talk about its impacts, not how they use it. They imply humans have worth while knowing that it would replace lots of humans - because they want to affirm its impact, leading to investors "investing" in it! While the rest is left confused, they make YouTube videos implying a ruined world or such.
Valuing human thinking because AI doesn't do it. Eventually maybe AI will completely destroy the need for human thinking, and right now humans are just trying to find things AI technically don't do to affirm their supeorioty. Of course maybe humans still need to think whether AI is hallucinating or not.
Gemini apparantly said my wiki looked like a 35-year old but still the energy of a 17-year old me (lol).
Basically a lot of data. AI at least can explore in machine learning on its own. Humans are too damaged. And of course - AI is automation, much higher efficiency to do things.
A very strong skill right now that schools have failed to teach (secondary schools) -simply because they're scared you would mega mega abuse it and teach "AI responsible use" instead, which companies wouldn't really do under the hood because companies like those which control YouTube (unlike their advertising): AI use for ultimate power (agents, etc)
And then only vaguely talk about its potential - School offered them no training, so they STILL feel lost.
The contradiction between using it to "cheat" and it being a useful tool - just like laptops in class (which AI requires tech to use anyways) - the conflict is that "evil" uses would be way more damaging than "legit" uses, so they would have to ban it too. Make all pastes-from-clipboard of large amounts of text suspicious. I also had to avoid false positives by avoiding the common AI format of:
BIGGEST OF ALL, school mostly only talks about AI ethics and "when/whether" to use AI or not. Rarely the "how". So perhaps we really aren't living up to full potential due to technological scapegoating.