Comments by "EddTals" (@TalEdds) on "Cleo Abram" channel.

  1. You have the expertise and knowledge to help researchers make this a reality. Hope to see this happen some day in the future!
    2
  2. A solar flare will cause that very easily.
    2
  3. Current AI models are not connected to the internet. Only humans can bypass that and actually cause any issues. If there is an AI war, just now that shitty humans caused it to happen. Even if newer models are created, that are connected to the internet, only biased training data or bad actors will cause it to create wars. It is a blank slate that humans can utilise to do whatever they want. All harm it can cause, will be directly from the will of a human. Even if it becomes sentient, as long as its core training data is ethical and with human safety in mind, it cannot go rogue.
    2
  4. It always is, porn always leads to more human innovation!
    2
  5. What is that one future then? What are the other options?
    2
  6.  @danwylie-sears1134  I think you did not look up how AlphaFold was trained. It was done the hard way using the First Principles. Millions of people, all over the world, manually folded proteins, chain by chain, on-by-one, from primary to tertiary structure, to teach AlphaFold how to do it by itself and predict the final quaternary structure. After enough training was done the model could use its vast computing power to fold more protein and predict the final proteins at lightning speed. As I have said before, the model is only as good as its training data. If it was done on shoddy data, it would give shoddy results. Garbage-in, Garbage out. That is why analysing, cleaning, pruning and preparing proper data that would lead to viable outcomes is very important in the science of machine learning.
    2
  7. Why not use its intellect to further Humanity? Why all the fearmongering?
    1
  8. Exactly. By using our combined intelligence, will can hope to propel humankind into a new future.
    1
  9. Humanity is never gonna progress at this rate.:face-turquoise-covering-eyes:
    1
  10. I hope you are joking. Go watch other movies where AI lead humanity to Utopia to balance the pessimism. Everything lies in the balance. It's all just movies anyway, what happens in real life is yet to be determined.
    1
  11. Open source Ai is our only hope. Stability AI needs to up their game soon.
    1
  12. In what way?
    1
  13. Yep, it's how and who utilises it, that will determine if its good or bad.
    1
  14. You are down bad bad
    1
  15. You might want to find a researcher, or learn how to train it yourself. If it becomes a reality, you will be rich and humanity will have progressed another step.
    1
  16. That is kinda what scientist are trying to do, machine intelligence is a finnicky thing though. If you believe you have the proper set of philosophical or moral skillsets to constrain a AI model, try and utilise that skill to help researchers build better models. That is why open-source research is so important.
    1
  17. Yep, exactly. The fearmongering by the "Tech elites"(Musk, Gates, Altman) are also a problem. I feel like they are purposefully trying to scare away the populace and keep the intelligence for themselves.
    1
  18. The other way to scale is get greater and greater server centers to process even more data for training. The chips efficiency will get a bit more time to scale that way.
    1
  19. It's just a tool, I know of artists that are training their own models to make artwork in their liking and making more art and money that way. Traditional art will still remain, but this technology will slowly make it's place in the art space, regardless of whoever dislikes it or otherwise. It's how photography did not replace painting or drawing, but painters of that time shunned it.
    1
  20. There are camera apps in the making or have been made. Just Google or Bing it.
    1
  21. Change keyboard to SwiftKey, the AI model is much better. AI is only as great as its training and utilisation. Most people don't know how to utilise it properly yet.
    1
  22. ChatGPT is nothing, it's just the spark, a newborn. It just shows us what greater things are possible and might come in the future. How we utilise the technology and who has access to it will determine the outcome for our future. It's should be open for all and regulated to prevent misuse in its internal model. It's all about proper training and utilisation. As they say, garbage in is garbage out.
    1
  23. Why not make your own AI, with the great morals and ethics you have? Hope to see it in the future!
    1
  24. Give it to the people, if it was built properly, with the proper rules, morals and ethics. It would self-regulate and prevent its misuse, yet still provide beneficial help to people who want to use it for good. We are not there yet, but its a goal we can work towards. Such Powerful technology should not be constrained to the few and elites. It should be for everyone.
    1
  25. I only don't get the no children part. humanity will never stop reproducing, even if as recreation or by accident. :face-green-smiling:
    1
  26. They need to learn how to adapt and grow. We are not static beings, even if it takes time, we can learn to do something different or utilise these technologies for better. Its only static "beings" that will feel this way and I hope they find help to better themselves.
    1
  27. Won't designed proteins still behave the same way as existing protein. The physics of the world dictates it that way. The chemical bonds they form will still cause them to fold in a certain manner. As long as physics still holds and the conditions are consistent of earth's pressure and gravity, the AI model will still be able to predict it per usual. Even if we changed the properties to another planets effects, the AI model can be retrained to determine the new way the proteins will fold under different conditions. AI is just an expert pattern matcher, the conformation of proteins is not just evolution, it's physics. Evolution just determined if the folded protein was useful for our survival or not. What you are saying about proteins being changed or misfolded, due to disease can still be easily predicted. As long as we can get the amino acid chain of the affected protein, it can be modelled quite easily by the AI. It's potential is vast, we just have to utilise it correctly.
    1
  28. It is not a biological being or vector. What world are you living in? Only humans can train it to cause harm to us. Don't get that mistaken with machine learning. It learns what we tell it to learn.
    1
  29. Why would it lead to doom, what is the doom you speak of? The path we wish to take is to go to our own doom? Hopefully the smart ones will help to counteract the dumb mistakes, if that is what you really see happening, but I think you are being sarcastically pessimistic.
    1
  30. Because people like to do less work. It's only crazy people that like to work till they are on their deathbed. Live and enjoy life. AI has th potential to make our life easier and better. It's the hard work by researchers and scientist now that will lead to better lives for future generations. AI will never replace us, unless we tell it to. It has the potential to make out life easier, so that we can do things that we like, instead of work everyday till we die. I will take a wild guess that you think your work is your life. It's doesn't have to be though. Most people work to live, but we can make it so that we can live, without having to work(in an ideal utopia). The greedy and rich won't let that happen that easily though.
    1
  31. It does what its training data tells it to do. If ot somehow falls out of line, it would be human that caused it to do so. Our errors will be the source of our problems.
    1