)'-.,_)‘-.,)'-.,_)'-.,)'-.,_)’-.,_

  • 4 Posts
  • 392 Comments
Joined 1 year ago
cake
Cake day: May 19th, 2024

help-circle
  • I like to think of it this way. Plugging in a USB-A connector is like observing a qubit in superposition—once you attempt it, the superposition collapses, and you instantly know if you were in the right orientation or if you need to flip to the opposite one!

    And just like conditional probabilities in quantum mechanics, the first try has a 30% chance of being right and a 70% chance of being wrong. But here’s the kicker: when the connector isn’t touching the port, it resets to a superposition, and the probabilities shift to 60% right and 30% wrong. That’s why even the third try isn’t guaranteed to be correct. As the number of attempts increases, the cumulative probability of eventually getting it right approaches 100%


  • Something very similar is also true with humans. People just love to have answers even if they aren’t entirely reliable or even true. Having just some answer seems to be more appealing than not having any answers at all. Why do you think people had weird beliefs about stars, rainbows, thunder etc.

    The way LLMs hallucinate is also a little weird. If you ask about quantum physics things, they actually can tell you that modern science doesn’t have a conclusive answer to your question. I guess that’s because other people have written articles about the very same question, and have pointed out that it’s still a topic of ongoing debate.

    If you ask about robot waitresses used in a particular restaurant, it will happily give you the wrong answer. Obviously, there’s not much data about that restaurant, let alone any academic debate, so I guess that’s also reflected in the answer.



  • Vibe coding works, but there are some serious caveats.

    I’ve used LLMs for data visualization and found them helpful for simple tasks, but they will always make serious mistakes with more complex prompts. While they understand syntax and functions well, they usually produce errors that require manual debugging. Vibe coding with LLMs works best if you’re an expert in your project and could write all of the code yourself but just can’t be bothered. Prepare to spend some time fixing the bugs, but it should still be faster than writing all of it yourself.

    If you’re not proficient in using a specific function the LLM generated, vibe coding becomes less effective because debugging can be time consuming. Relying on an LLM to troubleshoot its own code tends to lead to “fixes” that only spawn more errors. The key is to catch these situations early and avoid getting lured into any of the wild goose chases it offers.













  • Chemists write papers where precision is key, so they’ll use an IUPAC name like (RS)-N-methyl-1-phenylpropan-2-amine.

    The title of the paper will use the trivial name N-methylamphetamine, because it’s easier for everyone, but still precise enough.

    People who buy and sell the drug, use one of the trade names, such as: speed, ice, chalk, dunk, pookie, rotten candy, rocket fuel etc. at this point though, all precision goes out the window. The product usually contains a variable mixture of fun and interesting surprise compounds and even the concentration of the active ingredient can be all over the place.



  • Nah, we can do better than that. Just add microtransactions for premium responses, loot boxes for random fun personalities, and a battle pass for exclusive chat themes. Oh, and let’s not forget daily login bonuses for extra chat time, exclusive avatars for early subscribers, and a seasonal event that requires you to chat 12 hours a day to unlock the ultimate bot personality. Then we’re in business. And don’t worry, we’ll throw in some limited-time offers that expire in 10 minutes, just to keep things spicy! Plus, if you can maintain a streak of daily chats, you can earn badges and achievements. BTW which one sounds more exciting: gold coins or diamonds?