AI.

As a user of ChatGPT, Stable Diffusion and similar models my takeaway is mostly that these tools will enable a new paradigm of dialogue interfaces with computers, and will vastly improve our creativity, automate and filter through and visualize large amount of data and get insights, ability and offload computational brainpower, empower developers etc.

Its all about the iterative feedback loop between computer and human.

It might disrupt some jobs but it will improve peoples current output and also create new jobs. As this system is based on dialogue, and that is a critical component to this, there needs to be a human that takes the creative decisions off the output of the AI. Thr idea of AI running off building stuff on its own is still decades ahead, but we will see many interesting uses of this system that will make our tasks easier.

However the big concern for me is if the technology is gamed by governments in collision with big tech, training it towards their own end, Its vital that its open sourced or democratised and that the AI model is transparent and we as users have control over the configurations.

Reply to this note

Please Login to reply.

Discussion

I think that Altman is very naive if he thinks we can come into a constitutional agreement with UN what the guardrails are. If such thing comes to pass, the most ethical thing to do is to leak the software and data and make it public domain.

And another worry is if they tie the AI to a shitcoin or CBDC Ubi scheme like Worldcoin to get access. That can lead to some pretty dystopian scenarios.

Well, I don’t think we need to worry about that, since without a #mainvolume binary Intelligence license, artificial remains faked.

And being smart about it won’t result in a good outcome.

Our hope is an open-source competitor emerges.

The technology has the potential for amazing feats but also evil if it is a black-box and allowed to be controlled by govts and tech monopolies.

AFAIK, the one who concern about AI transparency mostly comes from AI researchers who are not restricted by some 'secret' grants from Governments or some Bug Tech. Those 'secret' grants will restrict the researchers to publish their code and AI model openly. The dilemma is the grants for big research in AI usually comes from Governments or Big Tech which have their own goal 🙂

Open Source AI research:

https://paperswithcode.com/

There may always be black box models, hopefully the open source ones can keep up. An important key is the data though. Imagine chatGPT + NSA raw signals data.