It's all fun and games to vibecode a clone of a an enterprise software tool for <$500.. until you actually have to make it enterprise-grade.
Great software for large enterprises hides a lot of complexity from users. Whether it's maintaining SSO， upgrading security / deps， RBAC， a wide gamut of config-gated features， API access， scaling， release management， integrations with changing APIs - there's a long tail of important but boring work to really make even simple apps work well in businesses. And it usually doesn't make sense for an Engg team to pay that recurring maintainence cost.
Companies continue to use SaaS tools like Linear， Notion， Slack， GitHub， Google Workspace， Figma， Salesforce， Workday， Glean and more.
There's a category of tooling this makes a lot of sense for： bespoke internal tools， dashboards and one-time / few-time use tools. A founder I know famously told me "I vibecoded my own data room tool and the folks I worked with actually really liked it" or "we vibecoded a better version of my Retool customer support UI" But for most other ones， it usually doesn't make sense.
One might argue that vibe coding ultimately reduces the cost to building competitors to such tools， and that there's some ensuing price compression. While there might be some truth to that， a version of Amdahl's Law still applies here. When software ceases to be the bottleneck， it becomes sales and marketing to compete. How many companies have moved off Slack since vibe coding？ How many Slack competitors have even come up since 2022？ Not many！
I imagine in the future， there will be a lot of cheaper less full-featured open-source alternatives for startups that will eventually try to compete. Hell， maybe models get so good that they pre-empt the need for all these features and one shot it perfectly out of the gate with a background loop that maintains it. Until then， "we'll just vibecode it" can be a trap.