Recent searches
Search options
I’ve been friends with @clive for a couple of decades now, and a fan of his writing (and thinking!) even before that. If you follow me and my work, I suspect that you’ll also get a lot out of this long, chewy interview with him about technology, thinking, mobility, infrastructure, the coming energy transition, writing, AI, and so much more: https://thegradientpub.substack.com/p/clive-thompson-tales-of-technology
@debcha @clive about 2 hours in, Clive starts talking about tools for talking to yourself so you understand what you are thinking.
That reminded me of a couple of years ago when @borthwick was looking for ideas around this, and I wrote https://www.kevinmarks.com/memex.html
And nots on the conference at https://www.kevinmarks.com/toolsforthinking.html
As one participant noted, the challenge is that this is mainly desired by journalists, autodidacts and students, who don't add up to a venture capital attracting market.
@debcha @clive @borthwick I still think there is something needed here. As X spirals into chaos, and Google no longer finds the thing you were thinking of, tools both for self discussion and peer conversation are needed more than ever. Do we rebuild blogs, Google Reader and Technorati?
Or is this what people like about ChatGPT etc - in many cases they're used as a rubber duck that talks back a bit, but is mainly reflecting your thoughts to you with affirmation.
@KevinMarks @debcha @borthwick
That's a great post, sir
And yeah, this is very much an unsolved problem ... and I've wondered the same thing about where exactly we're heading
The ChatGPT/rubber-duck thing probably obtains in some cases, but I would bet -- though I have zero data on this, pulling it riggght outta my hat lol -- the majority of folks are using it just as a google substitute
The people offering retreival-augmented tools for chatting with one's notes ...
@KevinMarks @debcha @borthwick
... are, I think, some of the few folks trying to work on this "tools for talking to yourself" area -- a bit, anyway
(It's certainly what Steven Johnson is taking a stab at with Google's NotebookLM)
Me, I'm a journalist, I can't put sensitive notes (*particularly* not interview transcripts) in the cloud, where they'd be used for training during Google or OpenAI etc's next *blood feast* for *fresh text*
But if we could get performant RAG working locally ...
@KevinMarks @debcha @borthwick
... ideally with models that are also much more energy performant, that could get interesting