Marc, another great post here. I'll expand your framing to say that regardless of our field, we'll need to find better ways to integrate generative AI into our workflows. You focus here on writing, but I can imagine that the same logic could apply for almost any profession.
Maybe the best example right now is the way that generative AI has been used for assisting with programming as in GitHub Copilot. Much of the ethos of your argument here applies there (e.g., helping to adapt the interface such that it aligns well with the way that these tools are most helpful in an existing workflow.) It's an interesting space to watch for sure and I can only imagine the number of tools that will emerge over the next year.
One of my biggest concerns is that it's going to become increasingly difficult to figure when and where generative AI tools have been used. It's already almost impossible to untangle. As you've mentioned various other places, the embedded and sometimes opaque ways that the algorithms make their way into the tools is a real issue that needs to be addressed.
Thanks for the update. I guess we will wait and see. Interesting to see how fees will work. Will we have to pay for both a LLM and an interface, or will they be bundled, etc.
Marc, another great post here. I'll expand your framing to say that regardless of our field, we'll need to find better ways to integrate generative AI into our workflows. You focus here on writing, but I can imagine that the same logic could apply for almost any profession.
Maybe the best example right now is the way that generative AI has been used for assisting with programming as in GitHub Copilot. Much of the ethos of your argument here applies there (e.g., helping to adapt the interface such that it aligns well with the way that these tools are most helpful in an existing workflow.) It's an interesting space to watch for sure and I can only imagine the number of tools that will emerge over the next year.
One of my biggest concerns is that it's going to become increasingly difficult to figure when and where generative AI tools have been used. It's already almost impossible to untangle. As you've mentioned various other places, the embedded and sometimes opaque ways that the algorithms make their way into the tools is a real issue that needs to be addressed.
I think the Copilot example is really apt. I know many of the CS faculty grapple with it and could certainly see other professions struggling as well.
Check out this podcast. This DJ is trying to design different AI musical interfaces for the same reasons we writers are looking for new AI writing interfaces: https://thegradientpub.substack.com/p/nao-tokui-ai-music-hci?r=2l25hp&utm_medium=ios&utm_campaign=audio-player
Are any of these tools available right now? This post really struck a chord.
No, not yet. They're all prototypes, but the devs are showing them as examples of what could be if we shift gears away from chatbot-based interfaces.
Thanks for the update. I guess we will wait and see. Interesting to see how fees will work. Will we have to pay for both a LLM and an interface, or will they be bundled, etc.