Can LLMs make idea-sharing truly frictionless?

I want to share valuable ideas on the internet but I don’t want that to be my main thing. I’d rather it be a byproduct of cool stuff that I’m doing.

The problem is that if I’m heads-down building, I struggle to casually synthesise my learnings for others. Maybe it’s that my brain’s not wired for this, but I need to spend atleast some time ‘maturing’ the ideas to communicate them more effectively and reach people.

Until now, I’ve typically alternated between “building” phases (where I’m focussed on learning, doing projects, writing code) and “output” phases (where I’m focussed on distilling ideas and sharing them).

A “building” phase might last a year or two. During the phase, I might get a bunch of ideas but then never get round to sharing them. By the time I come round to an “output” phase, they’re forgotten (or I’ve internalised them and no longer able to express what I once could have).

Is there a way in which I can still share ideas (relatively) effortlessly, as I go along, even if heads-down focussed on building?

Lowering the friction

I’ve played around with trying to make sharing ideas easier in the past:

  1. Setting up automation: I wrote code which takes anything I write in my personal text editor (Obsidian) and, with the addition of a single tag, will automatically share it on my personal website.

    a. Problem: This helps, but it doesn’t solve the empty-sheet-of-paper problem.

  2. Better scheduling: Setting aside time each week to synthesise ideas and share them.

    a. Problem: it’s very easy to end up using this time for other more immediate priorities.

  3. Using LLMs to redraft my articlesa

    a. Problem: Just using LLMs to edit / redraft doesn’t get you all the way there - you still need to put in the hard work to synthesise the ideas well.

However, I think we’re now at the point where we can make the friction to sharing ideas so low, that it may actually become sustainable to do alongside true deep building. I’m envisioning a system that allows something like:

  • store super rough thoughts - whether writing in a daily journal, recording a quick audio note while I’m driving, making a quick video, maybe even ‘ambient’ recording (in the background)?

  • an LLM automatically “tidies” and “distills” those thoughts

  • it then converts them into the best ‘native’ formats to be shared on different platforms - e.g. tweet threads for twitter, articles for a personal blog, etc. (In the future, maybe even create audio / video content.)

Maybe there’s scope for a back-and-forth between the LLM during the creation process. The LLM could prompt you to think about certain recent experiences or learnings to get the ball rolling - and it could ask clarifying questions to help it distill the thoughts effectively.

I’ve seen Jason Liu (a fellow builder in the LLM space) do some variation of this, for example. He has a bunch of blog posts that are transcriptions of audio recordings (eg), as well as tweets which are transcriptions from him talking into a camera (eg).

So if this works, maybe my process simply becomes: Spend the day doing cool stuff. Sit down for 10 mins at the end of the day, answering questions from an AI and riffing into a camera. And all the rest is taken care of.