I keep seeing more signs that Google will have to get its house in order to stay competitive. Like this one [https://archive.ph/7BYDG]. When that happens, many free or very cheap services will simply be discontinued. Let’s not forget that “Google graveyard” grows every day.
Microsoft’s purchase of 49% of OpenAI (it already had a prior collaboration agreement, whose fruits we’ve seen with GitHub Copilot). If Microsoft effectively integrates OpenAI’s tools into its SaaS like Microsoft Office, many of Google’s “customers” will migrate to its services. That includes Bing (Bing+GPT-4) and Office, as paid SaaS services. Professional users will gladly pay.
Google only knows how to make money by selling ads. It sells SaaS, but it’s not its core business nor part of its culture. The poor support they offer proves it. Has anyone in the room managed to talk to a human at a Google service? And a ChatGPT‑like assistant that substitutes for a search engine doesn’t show ads, and therefore doesn’t make money under Google’s current business model. How does Google solve that?
Where does that leave the continuity of Google Apps (Drive, Classroom, Meet, Calendar, even Gmail)? Very likely, over the next 5 years these tools will evolve more than in the last 25 as generative AI gets integrated.
Keeping these tools state of the art and footing the compute bill (the GPUs that run AI don’t have zero marginal cost; they cost money) will be a business decision. And we have no certainty that Google will choose to keep playing.
And even if the probability of discontinuing the service, or letting it go stale as a low value‑add free service for “free” users, is very low, that’s the risk every organization takes when it uses them for critical processes.
Organizations like the UPC are taking on significant risk and technical debt. When it would be easy for us to self‑provision with open‑source solutions, keep our IT services staff skilled up, and create business opportunities for UPCnet or inLabFIB.
In the article Privacy and E-learning a Pending Task (Alier M, Casañ Guerrero MJ, Amo D, Severance C, Fonseca D. Privacy and E-Learning: A Pending Task. Sustainability. 2021; 13(16):9206. https://doi.org/10.3390/su13169206) we already discussed the long-term effects of outsourcing core services at universities. In that case, the bet on Moodle and Open Source proved to be the right one back in 2003. What’s going to happen in 2023?