3 ms·
We are at the precipice of epoch defining technology with GPT-4. But it's being held back by our most incapable... It's like if we wanted to stop people havin
by psychphysic 3y ago
We are at the precipice of epoch defining technology with GPT-4.
But it's being held back by our most incapable...
It's like if we wanted to stop people having access to calculators because some were misusing them for their tax returns.
- Decabytes 3y agoI think this mischaracterizes the whole issue. There are already Open Source LLMs that can be run locally, You can use OpenAIs API which gives you access to multiple different types of models that can be trained for your use case. People keep creating Jailbreaks to get around restrictions put in place by the developers of various models etc. The only thing that is holding back AI is the resources required. Most people can't to afford train a LLM by themselves from scratch, so it's not a viable avenue for widespread ai use. Also you can use GPT 4 if you subscribe to ChatGPT plus or pay to use it's API
- psychphysic 3y agoI disagree, the 3-4 versions of GPT-4 I've used have face planted in terms of usefulness. No doubt in my mind this is to do with RLHF. It maybe be cost effectiveness of running the models sure. I'm trying to be open minded.
- simonw 3y agoFrom the article: "I don’t think cases like these invalidate the potential of ChatGPT and other chatbots. In the right scenario and with the right safeguards, it’s clear these tools can be fantastically useful." It's not calling for people to lose access to these tools.