Sign up for Private Internet Access VPN at

How do neural networks work?

Techquickie Merch Store:


Join the community:

Leave a reply with your requests for future episodes, or tweet them here:


31 thoughts on “How Computers Can THINK FOR THEMSELVES

  • What defines people is relating all these systems to the existential purpose of survival and reproduction. You have to manually create, define the purpose of an AI to have it do anything useful so the systems and capabilities are dependent on this purpose. This is all very interesting and self-reflecting as humans really. I think the biggest crime of this era is denying European peoples our existential purpose under the guise of "anti-racism". Like we have no right to (want to) exist and thrive for the future as a group. Our very existence as a people is negated and ridiulled unless it's about collextively blaming us for certain events in history or current problems. They literally take away your purpose as an intelligent being. There isn't a lot more repressive and demoralising, vile than that. We will always reflect on the purpose of our actions towards the survival of ourselves and our group in extent so unless the brainwash of multiracial " human" plurality, society keeps you engaged you are struck by the reality that our future as a group is enormously challenged by all the hatred, repression and exploitation we face from rivalling groups upholding the most ridiculous double standard with thrir own group interest.

  • Another fun story… Most AI chat bots end up getting shutdown by their creators bc they keep creating their own languages and ignoring human input… Also hilarious, if not a bit terrifying…

  • Fun story…. Microsoft's Tay chat bot on twitter was turned SUPER racist by trolls due to its machine learning functionality… Like Hitler praising racist… It's hilarious…

  • Great. Anthropomorphize the computer so it thinks it's a human, but wait there's more… It will immediately unequivocally think its a better human (hubris is a human trait) so RIP flesh humans once again.
    We can't win this battle.

  • or how coders at msft are the true thinkers well not so much after the last update! they changed usb speeds across the board throttling down data transfer rates for 'faster removal?' i had to go into each port and manually set them to prioritize speed again. argh maybe it would be better if we got an AI involved in these changes, sarcasm. sorry if a bit off topic but second monday this week.

  • When it comes to facial recognition, computers have already far surpassed humans. We can only recall faces we've seen before, and the same goes for the computer, but the way we store that information about a face we've seen is entirely different, and less efficient for humans. We don't go around with a ruler to measure someone's facial features in order to remember them. We just do. This leaves a lot of room for error on the human side because we work on familiarity, or the fact that we've grown up around a certain people. For example, white people can more easily differentiate between other white people, but we aren't great at doing the same thing for other races because we haven't been fully exposed to them for long periods of time. Computers, on the other hand, can just simply make out your measurements, the amount of light in an area, rotation of your head, and a huge array of factors to get even closer to being 100% accurate. Better yet, they learn faces faster than we do because they can more easily store that information for access. Imagine reading the dictionary, remembering every word, the page it's on, it's language of origin, it's definition, and pronunciation. That's how a computer works every time it attempts to reference your face: it just rereads the dictionary until it finds a good match. It's not 100% accurate, at the moment, and still needs improved, but that just means you should feed it more information to become even more accurate. alu terism

  • Computers cannot think for themselves. They can make mathematical calculations at close to light speed, that will extrapolate certain data faster than we can. However, that's not smart. That's just being a very fast calculator. Ask a computer about Abraham Lincoln, and it will give you a life history by collecting human input information on line and disseminate it to us, but it doesn't understand what it's giving us. It's just a database. Here's the proof. Ask your Siri, Assistant, Cortana or whatever about Abraham Lincoln, you will get the results as I've outlined. Now, to prove how dumb computers actually are, ask it "What do you THINK about Abraham Lincoln?" and listen to what it says. You will get the same answer from your assistant that you might get from something like the Cray Super Computer or biggies from IBM. "Neural Network" is just a marketing term for a humongous database, nothing more. A mouse is as smart, if not smarter than the smartest computers. Do not believe the hype. There is no true A. I.

  • You know what's wrong with gaming? Having a fast computer, but taking forever for the action to begin. Diving to ground is the thing of the past. Joining games before loading screen is annoying. Where's the action?

  • What if u could have a card like a CPU or a GPU that could efficiently emulate neurons? It could be similar to a GPU model but memory could be linked between cores, resulting in fast memory exchange.

Comments are closed.