• 0 Posts
  • 9 Comments
Joined 1 year ago
cake
Cake day: June 12th, 2023

help-circle


  • The thing that shits me about this is google appear to the public to be late to the party but the reality is they DID put safety before profit when it came to AI. The sheer amount of research and papers put out by them on AI should have proven to people they know what they’re doing.

    And then openAI throw caution into the wind and essentially make google and others panic knee jerk because there’s really money to be made, and now everyone seems to be throwing caution into the wind and pushing it into the mainstream before society is ready.

    All in the name of shareholders.







  • It’s a technical set of problems.

    • Where does the front end live, who hosts it, who pays for the compute, and what determines the latest version is the truest version
    • where does the data sit. Posts, media, content, and how does this get referenced properly in a safe way.

    What’s funny is truely distributed compute is totally possible today, thanks to a lot of work done in the blockchain community. Notice I said blockchain and not crypto, we don’t want the bullshit associated with that (coins, nfts etc). What we want is distributed compute and storage that can be read in a way that provides the same function as Reddit etc. Coupled with a good client experience like sync.

    The biggest problem with that though is that blockchain that is truely distributed is slow by nature, because each block of data is distributed and validated to all nodes that host to keep consistency. And the larger a site becomes, the more data there is to store, and the more resource intensive verification becomes so therefore the nodes slowly gain a higher set of requirements.

    So the middle ground is something like Lemmy. Where you can run your own instance, that talks to a wider federated network of instances where no one single entity can control the content.

    In tech, a lot of the above is explained by a concept called CAP theorem. It’s a really interesting problem that has only really been solved by a few vendors (google spanner is a good one) but even then it doesn’t cover the distributed part.