The man who killed Google Search?

Ex-Google search engineer here (2019-2023). I know a lot of the veteran engineers were upset when Ben Gomes got shunted off. Probably the bigger change, from what I've heard, was losing Amit Singhal who led Search until 2016. Amit fought against creeping complexity. There is a semi-famous internal document he wrote where he argued against the other search leads that Google should use less machine-learning, or at least contain it as much as possible, so that ranking stays debuggable and understandable by human search engineers. My impression is that since he left complexity exploded, with every team launching as many deep learning projects as they can (just like every other large tech company has).

The problem though, is the older systems had obvious problems, while the newer systems have hidden bugs and conceptual issues which often don't show up in the metrics, and which compound over time as more complexity is layered on. For example: I found an off by 1 error deep in a formula from an old launch that has been reordering top results for 15% of queries since 2015. I handed it off when I left but have no idea whether anyone actually fixed it or not.

I wrote up all of the search bugs I was aware of in an internal document called "second page navboost", so if anyone working on search at Google reads this and needs a launch go check it out.

> where he argued against the other search leads that Google should use less machine-learning

This better echoes my personal experience with the decline of Google search than TFA: it seems to be connected to the increasing use of ML in that the more of it Google put in, the worse the results I got were.

It's also a good lesson for the new AI cycle we're in now. Often inserting ML subsystems into your broader system just makes it go from "deterministically but fixably bad" to "mysteriously and unfixably bad".

I think that’ll define the industry for the coming decades. I used to work in machine translation and it was the same. The older rules-based engines that were carefully crafted by humans worked well on the test suite and if a new case was found, a human could fix it. When machine learning came on the scene, more “impressive” models that were built quicker came out - but when a translation was bad no one knew how to fix it other than retraining and crossing one’s fingers.

Yes, but I think the other lesson might be that those black box machine translations have ended up being more valuable? It sucks when things don't always work, but that is also kind of life and if the AI version worked more often that is usually ok (as long as the occasional failures aren't so catastrophic as to ruin everything)

> Yes, but I think the other lesson might be that those black box machine translations have ended up being more valuable?

The key difference is how tolerant the specific use case is of a probably-correct answer.

The things recent-AI excels at now (generative, translation, etc.) are very tolerant of "usually correct." If a model can do more, and is right most of the time, then it's more valuable.

There are many other types of use cases, though.