I’m are exploring an app idea built on Autonomi and are curious how others here look at it, and whether there are technical challenges we might be overlooking.
The concept is straightforward:
Users upload files to Autonomi and can then “spot” them in an application. A post would contain:
Title
Description
Optional image
Category, for example video, audio, or data
Optional subcategory, for example film, series, or documentary
The data address (content string)
When opening the application, the client downloads the index of posts, caches it locally, and shows all new posts so users can browse them. A download button would trigger the client in the background to download the file using the ant client.
For an MVP we think this could be very simple:
Posting without accounts
A single index uploaded to Autonomi
The client downloads and caches the index on startup
User can set download folder
User can adjust download speed, for example chunk batch size
Possible later improvements:
Posting or “spotting” using wallet signing, for example paying a small fee to be included in the index
Moderation, ideally decentralized, to reduce spam or illegal content
My feeling is that something like this might be possible to build in about a week and could already deliver useful value.
Questions:
How do others here look at this idea?
Has anyone already experimented with something similar?
Are there technical complexities around indexing, discovery, or scaling that we should take into account early?
Sounds like you’re defining a sort of personal search engine, where you tag and add them to your index, so you can search them later etc?
It is hard to fully understand the concept. If it will only take you a week, why not throw together a prototype and see how it flies?
I think search is a hard problem to solve in a distributed network. Certainly, in the way we think about it now.
Maybe we need to be thinking about personalised crawlers, perhapa using AI to categorise data to your style, then use it to index a database accordingly?
If multiple databases could be easily cross referenced, with personal overlays supported (i.e. your agent’s take on it extends it), that would be pretty cool.
Having an append only structure would probably be necessary for performance too, so you could easily update your local copy from your favourite remote sources (maybe, one day, google would be a source too, etc).
My view on this and why I bailed on the Atlas app noted by @aatonnomicc and @DavidMc0 is that having no control of what others shared via it started concerning me.
Whether or not I ended up liable or not starting becoming a stress I am not interested in living with.
Be sure that you are ok with that.
There are may arguments as to why you won’t be, you only need that 1% chance of illegal or copyright being linked via your app, even if you are not liable there is still a non zero chance of huge legal costs in making that case.
I think most folks stopped after the IF money dried up. There was some bitterness with how there wasn’t more support from Maidsafe too, I gather. Some of us are still using our spare time too create stuff, but it is a lenghty process, even with AI help.
I personally think search will be user supplied indexes to a “search engine”, whether that be some sort of distributed app like what this proposal seems to include or it be some person/group more centralised database kept on autonomi and people will pick and choose the one they want to use, and even multiple.
The old idea of crawling will have limited use since links may not exist for a lot of files even though they are public. The submitted indexes would have to be public themselves and likely be part of a protocol where the primary record is findable with ease by the search engines building their databases.
Yes, it is a shame how little support followed Impossible Futures.
While it’s hard to actually develop stuff, to me the lack of viable concepts is more concerning than the lack of functioning apps.
Of course the tech side needs to progress, but nobody seems to be creating compelling concepts that would add value to existing markets / products & laying out the business & marketing case for them, their value add & why adoption is likely if built & marketed.
Once compelling product concepts are developed, the challenge becomes getting the tech to make them happen, but making tech without product-market-fit won’t pull through the adoption the network will need to grow.
I’ve been toying with organising an Autonomi ‘product concept’ competition to inspire people to put forward app ideas with clearly defined value-add to specific addressable markets. I may give it a shot following the Autonomi 2.0 announcements if it seems appropriate.