Gregory's ServerWhenever anyone says "Look at #ChatGPT now and imagine what #AI will do in three years", I only smirk. It reminds me of my teenage years, where people said the same things about #Crypto, #VR, and #AR. And yet, I still fail to see this technology become a meaningful, if at all, part of my and others' lifes. Yes, the AI models of today are impressive beyond belief, but the stuff they generate is still too imperfect for me to use them.
 7 comments
Since the release of #ChatGPT, I have been trying to feed it questions that I have in my day-to-day college and work life. Amongst hundreds of prompts, it managed to help me _once_. GitHub #Copilot is better for coding, but it's only good for writing repetitive code after one has written the first line of it, and even then is it imperfect. Most of the time it is just plain irritating, so I keep it turned off. @kytta I guess it depends on your area of work, it's helped me quite a bit with my Google cloud learning (in my spare time) It is hard to predict what it can do in 3 years. 10 years ago, we were just 2 years away from full self driving cars. Turned out the last bits are quiet hard. However, grounding an argument in "I do not use it so I doubt it is/will be useful in general" is hard to defend. The question could be "can you imagine use cases where it is useful today?" and how many cases could you make? @kytta I agree with you in that, the problem is that, even if those thing didn't become part of our lives, they made rich people richer (and maybe even poor people poorer) without them paying anything. Take crypto for example, millionaires invested in it without paying taxes for a long time (and they are still doing it maybe) The same goes for AI, with monsters like OpenAI not even giving credit to works they stole to train this tools @kytta I‘m teaching statistics and tried to get ChatGPT to create a useful exercise cheet for my students. No chance. It‘s just creates stuff that is mathematically wrong. It can‘t even for once create a correct solution without breaking basic rules. In the end, it might not be wise to train an AI on random information that might or might not be correct. It‘s like the old saying: garbage in, garbage out. |
Yes, I am sceptical, and I may be wrong. But for now, I am not worried about the AI taking all our jobs away from us, but rather about the bulls**t it spits out that other believe 🙄