this post was submitted on 17 Feb 2024
85 points (75.4% liked)
Technology
60102 readers
3134 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
There are lots of things that are very hard to program, but people can do very easily. For example, play Go or recognize that an animal is a bird.
Machine learning/ai makes it competitively simple to make computers do some of these things, but at the cost of efficiency and speed at runtime. This is true if computers vs people as well, a human brain is much slower, less efficient, and less accurate than a calculator.
Machine learning/AI is exciting because it enables computers to quickly be trained to do tasks that were impossible or would have required years of dedicated effort. The tech world is excited about it because whole new enterprises and areas of tech may spring up, big markets that were previously out of reach.
Downsides:
AI uses a lot more electricity. Especially for things that computers can already do, using AI is very inefficient.
Limited control. You train an ai model to do a task, but you don't have direct control over how it thinks. If chatgpt gives a wrong answer, they can't just trace the program and figure out why. It takes serious effort to figure out how chatgpt answers simple questions, so figuring out how it gets complex answers or why an answer is wrong is nearly impossible at this point. This also applies to unwanted behaviors,if you had a really good history chatbot who happened to turn out racist, you can't just turn that off. You end up having to retrain the model, or secretly add "make sure your answer isn't racist" to every submitted prompt.