gemini://jsreed5.org/log/2023/202302/20230201-gemini-and-google.gmi

Michael Nordmeyer asked what other Geminauts would think if Google added support for Gemini and began indexing capsules^. Three possible scenarios were presented, in order of most to least preferred:

Google provides SERPs for Gemini pages on the Web using HTML pages and only provides links to the actual content.

Google provides SERPs for Gemini pages on the Web using HTML pages and serves the content over a Web proxy.

My guess is that out of these three scenarios, Google would be least like to implement the first, precisely because it's the most difficult to monetize. I see them playing out the second scenario to test their ability to scrape Geminispace, and were they to implement Gemini search as a full feature, they would probably opt for the third scenario, at least to start with.

The second scenario is most likely to me if Google were ever to implement Gemini support in Chrome. While implementing such a feature would surely increase engagement with Gemini by several orders of magnitude, it would not be a healthy phenomenon for Geminispace as it current exists. More people--especially more people who treat the Internet flippantly--means more scammers and spammers, less meaningful and productive discourse, more bot abuse, greater potential for organized harassment or DDOS attacks, and massive strain on an infrastructure largely built on small, independent capsules.

Gemini support in Chrome presents another danger: Google being able to control the direction of the protocol, simply by its sheer influence. Were Chrome to add support for Gemini, it would switch from having 0 percent of the browser share to 99-plus percent of the browser share overnight. Google could easily extend the protocol to add its own proprietary features, create vendor lock-in, or abuse loopholes to monetize and track users. I've talked before about how a malicious actor doesn't even need to stray outside the existing spec to do this^^. A company like Google, filled with some of the most inventive minds in the world, would surely find ways to commercialize Gemini that most of us could never have imagined.

I should mention that Google already indexes some Gemini content via existing third-party Web proxies. Searching for "jsreed5.org" on Google yields some proxied links to my log, my twtxt file (which is only served over Gemini), and a few other pages. This is not a good situation; if Google were indexing these pages directly, they would presumably obey any robots.txt rules implemented by the capsule, but since third-party proxies are not crawlers and thus ignore robots.txt files, Google likely indexes undesirable content such as plaintext contact information.

Ultimately, my desire is for Google to leave Gemini alone entirely. The good news is that I don't realistically see Google having an interest in Gemini anytime soon. Gopher has been around for decades and is experiencing a resurgence, but Google isn't exactly champing at the bit to (re)introduce support for it.

Gemini and Google

2023-02-01