home.social

#softwareeconomics — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #softwareeconomics, aggregated by home.social.

  1. @dangillmor Excellent points, strong agreement and shared concerns. I was just discussing this with friends a few days ago.

    The proprietary software market has always had the problem that marginal costs drive unit revenues down where direct sales are at issue and that market size and corresponding lock-in factors are sufficiently valuable of themselves (future sales + cutting off the air supply of other proprietary-software competition) that companies would actively seek out lower prices for higher market share. E.g.,

    "Fighting China's Pirates" (2010)
    online.wsj.com/article/SB10001

    Several companies have long preferred leasing or subscription based models, most famously and originally IBM, also major enterprise vendors (Oracle, Peoplesoft (RIP), SAP, SAS, Salesforce, etc.). Apple's hardware focus (increasingly supplemented by entertainment subscriptions) is another.

    Sprinking LLM AI pixie dust over software makes the licensing / subscription model all the more viable, with additional moats of the lock-in afforded by a proprietary LLM model and the immense costs of developing and training AIs (see Microsoft's multi-billion dollar investment in OpenAI, largely in the form of Azure Cloud credits).

    Then there is the data access issue (dust-up yesterday on HN involving DropBox who claim rights to customer data for AI training: news.ycombinator.com/item?id=3 source: arstechnica.com/information-te. Principle discussion: news.ycombinator.com/item?id=3).

    This does make the Free Software world all the more attractive. That's been my preferred model for decades now. Question is whether or not AI/LLM actually does provide a sufficient use-case advantage over unassisted software. That's ... going to be an interesting situation watch.

    #ai #llm #SoftwareEconomics #Moats #Monopoly #FreeSoftware #FOSS #Privacy #Trust