Comment by idle_zealot

Comment by idle_zealot 9 hours ago

11 replies

View on Hacker News

Was the hard part ever really the software, though? It's the Service part of SaaS that seems to provide the moat. Lock-in, habits, workflows, integrations, and trust. And don't discount the appeal of making some part of your operations "someone else's problem." Could you hire engineers or use an LLM to make your own Google Docs? Probably, yeah, but would that be worth the headache of being responsible for a bespoke internal document system?

jonathaneunice 9 hours ago

You might think you can, for a while. Been there, done that. But you probably can not do so sustainably in most cases. Even if you could, would you really be better off building vs. buying? Outsourcing development, operations, and maintenance is almost always the better choice, letting you focus on the things you do uniquely, differentiably, or meaningfully better.

"We have this awesome internal version of Docs that we're responsible for fixing, upgrading, and doing support for" is not the flex "AI can code anything!" aficionados think it is. Especially when you also have similar internal versions of Sheets, Jira, Slack, GitHub, Linux, Postgres, and 100 other tools.

Reply View 0 replies
bryanlarsen 9 hours ago

The article is about SAP, Salesforce, etc.

Making your own Google Docs is stupid unless your company's core business is document management.

OTOH Replacing SAP with a bespoke system will make a lot of sense for many companies.

SAP is already the worst of both worlds. It'll have been highly customized for your flow so you've got all of the headaches of bespoke software and all of the headaches of SaaS. And unlike Google Docs, it'll be highly integral to your core business.

Reply View 5 replies
  • cedilla 9 hours ago

    Companies pay millions and millions to get away from bespoke software, but not simply because of the costs. Companies want to do their core business, they don't want to also be a software enterprise, and assume all the risks that entails. Even if AI makes creating software 10 times less expensive, that doesn't really change.

    Reply View 1 reply
  • louiereederson 7 hours ago

    you are aware of the long history of organizations being absolutely screwed by bad erp implementations right? nike's 2001 issue, the horrific birmingham oracle implementation, avon, etc.

    Reply View 1 reply
    • bryanlarsen 7 hours ago

      Those are examples of very good reasons to ditch i2, Oracle, SAP, et cetera.

      Reply View 0 replies
  • calvinmorrison 8 hours ago

    what do you mean SAP? like the ERP system?

    I would absolutely NEVER steal or rewrite that. So much finanical stuff is baked into the business logic that impacts finance, regulations, hr, etc.

    No do not roll your own ERP core.

    Roll everything else

    Reply View 0 replies
airstrike 9 hours ago

the problem is an AI can figure out habits and workflows pretty seamlessly. lock-in is artificial and loses power when it's really easy to make a competing app for large swaths of web apps.

integration is likely the most valuable part of the puzzle, but it's also prone to disruption

I think all that's left are like <50 apps each with their own very bespoke and "power user"-ready interface

Reply View 0 replies
llmslave 9 hours ago

yes, if it takes one month to build something that took 9 months previously, it completely changes your go to market strategy

Reply View 0 replies
senko 9 hours ago

> Could you hire engineers or use an LLM to make your own Google Docs

Or you can just ask your LLM to install https://github.com/CollaboraOnline/online

Between open source, LLMs, and SaaS vendors getting greedy and privacy invasive, the total pain minimization calc might shift for some orgs.

Reply View 1 reply
  • idle_zealot 9 hours ago

    Even then, I would expect most orgs would want to contract out to a company that manages an instance of that open source software. That management company could undercut bigger players because they don't need as many engineers working on features. I don't see where the LLM comes in and shifts the calculus here.

    Reply View 0 replies