Such a good point about how the comments get lost and overwhelming while editing in a tool we love! I look forward to learning about your 3 beginning topics.
Looking forward! Additionally, I would love to hear your perspective on the future evolution of data infrastructure in conjunction with advancements in AI [1][2].
That's an interesting topic, but not one I think about these days. Certainly there will be a lot of value in making the right information available to the right AI in the right format at the right time, and just by saying it that way it becomes obvious that there will be ample room for innovation. But I haven't been following specific techniques.
Great post—have you checked out Audrey Tang’s work? Former Minister of Digital Affairs in Taiwan and the tech tools they use to promote civic engagement, survey population for agenda items for regulatory discussion, prepare electorate for disinformation dump by nefarious forces…made me very optimistic for the future pro-social engagement. Here’s a good intro if you haven’t heard of her https://podcasts.apple.com/ca/podcast/machines-like-us/id1484910273?i=1000663786733
Honestly this sort of tool always leaves me scratching my head. I appreciate that some people find it to be valuable, but I'm not sure it gets at the issues I'm getting at here: keeping different aspects of a discussion in sync, reconciling new information, etc.
It seems we all want somewhat different things from notetaking. In addition to curating my external Brain, I would love to have conversation threads connected to different "nuggets" of thinking, ways of comparing notes across tools and more.
Your “hero” sounds like a combination secretary and president. Those were traditionally separate positions, so it’s no wonder it becomes an enormous amount of work for one.
What if agentic AI can soon *become* the "Project Hero"? 🤔 (at the very least in terms of keeping things in sync and updated, and summarizing what humans still need to contribute) Obviously there is risk and the need to earn trust, which might be a difficult problem to overcome for serious tests, much less full deployment. But it seems increasingly possible within the next 6-12 months, no?
Yes, presumably this will happen over time, and it should be great.
As to *when* it will happen... that's a tricky question. It depends on how quickly model capabilities advance, how good a job application developers do of building "project hero" tools around those capabilities, and the extent to which people are ready to adapt still-imperfect tools. I'd hesitate to guess. I would put (a little) money on the problem not being fully solved in 12 months, but I would expect interesting progress – shipping products that provide helpful assistance – in that timeframe (I presume it's already been nibbled away at today, though I haven't followed any specific developments).
Such a good point about how the comments get lost and overwhelming while editing in a tool we love! I look forward to learning about your 3 beginning topics.
Looking forward! Additionally, I would love to hear your perspective on the future evolution of data infrastructure in conjunction with advancements in AI [1][2].
[1] https://techcrunch.com/2024/06/21/openai-buys-rockset-to-bolster-its-enterprise-ai/
[2] https://www.cnbc.com/2024/06/04/databricks-is-buying-data-optimization-startup-tabular.html
That's an interesting topic, but not one I think about these days. Certainly there will be a lot of value in making the right information available to the right AI in the right format at the right time, and just by saying it that way it becomes obvious that there will be ample room for innovation. But I haven't been following specific techniques.
Great post—have you checked out Audrey Tang’s work? Former Minister of Digital Affairs in Taiwan and the tech tools they use to promote civic engagement, survey population for agenda items for regulatory discussion, prepare electorate for disinformation dump by nefarious forces…made me very optimistic for the future pro-social engagement. Here’s a good intro if you haven’t heard of her https://podcasts.apple.com/ca/podcast/machines-like-us/id1484910273?i=1000663786733
I know of her and gather she's doing great work but have not explored in detail. Thanks for the nudge!
Is this what some of the notetaking might look like? https://bra.in/9pKgKB
Honestly this sort of tool always leaves me scratching my head. I appreciate that some people find it to be valuable, but I'm not sure it gets at the issues I'm getting at here: keeping different aspects of a discussion in sync, reconciling new information, etc.
I did say "some" of the notetaking :)
It seems we all want somewhat different things from notetaking. In addition to curating my external Brain, I would love to have conversation threads connected to different "nuggets" of thinking, ways of comparing notes across tools and more.
Your “hero” sounds like a combination secretary and president. Those were traditionally separate positions, so it’s no wonder it becomes an enormous amount of work for one.
What if agentic AI can soon *become* the "Project Hero"? 🤔 (at the very least in terms of keeping things in sync and updated, and summarizing what humans still need to contribute) Obviously there is risk and the need to earn trust, which might be a difficult problem to overcome for serious tests, much less full deployment. But it seems increasingly possible within the next 6-12 months, no?
Yes, presumably this will happen over time, and it should be great.
As to *when* it will happen... that's a tricky question. It depends on how quickly model capabilities advance, how good a job application developers do of building "project hero" tools around those capabilities, and the extent to which people are ready to adapt still-imperfect tools. I'd hesitate to guess. I would put (a little) money on the problem not being fully solved in 12 months, but I would expect interesting progress – shipping products that provide helpful assistance – in that timeframe (I presume it's already been nibbled away at today, though I haven't followed any specific developments).