Hacker News Reader

Comment by Aissen | Hacker News Reader

Actually, there's another alternative. Provide all these "insights" with offline, data-on-computers-you-own systems. The intelligence can be carried with you, without internet. This is traditional computing. Apple or Microsoft, or even any other competitor could provide that. They just need to sell you this software (oh and also build it first).

View on HN

Not possible. The insights come from the combination of everyone's data, not the analysis of any individual in isolation.

I am sure there is still value in one users data that could be used that way.

I'd like to disagree. When Google Now reads my email to track my delivery package number or flight number, it's my data. When it understands that I like sport team X, it's my data. When it knows where I work and live to ask Google servers the weather or how long it would take to go somewhere, still my data used as the request base.

When Google(+) Photos auto-tags my photos, still my data, although I'd admit the IA training requires a big amount of photos.

Looking at https://www.google.com/landing/now/#cards ; it's obvious a lot of those need online information. It's not obvious that the data source for requiring those online informations has to be in the cloud, or combined from multiple people's data.

Right, It's your data because it's the things you care about, but knowing you isn't what google cares about, they want to know where you intersect with everyone else.

I think your point is good, and depends on the question of how much can be done on current and future hardware, and what differences might result to product quality in local-only storage and processing.

there is a point though in having as much data as possible for trainig and learning purpose.

Even with a locally installed AI, it will need to be trained on real and up to date data, it needs feedback on what was well recognized and on which data it failed, etc.

the training and update cycle can be done on an individual basis, but having a distributed system sharing a huge set of varied data would accelerate the process exponentially.

Ideally training data shouldn't be stored once it is ingested by yhe AI, but we'd need at least interconnected AIs dharing data from different users

Agree to a certain point. When Google wants to make suggestions based on preferences of other people (known as collaborative filtering), how do you suppose that information is computed?

>When it understands that I like sport team X, it's my data

It is, but that's pretty useless unless Google also knows so that it can send you the relevant updates.

>When it knows where I work and live to ask Google servers the weather or how long it would take to go somewhere, still my data used as the request base.

The weather, possibly (although again, it's not much use unless Google know enough to target the local weather at you). The how long to get there, not so much. One of the key things they use is info on where everyone else is. This is a classic case of pooling everyone's data in order to add value.

I agree about the value Maps/Waze style pooling of information. Also agreed that social (what your contacts like) recommendations might be useful. Leaving aside the possibility of implementing those in a P2P fashion, I'm just saying this is not all there is, and that is in fact a minority of the useful stuff.

It is, but that's pretty useless unless Google also knows so that it can send you the relevant updates.

It's shortsighted to say that just because (example implementation) my device polls their server for $team, that it is the same as "google knowing" I like $team. The interface does not need to be authenticated, or keep logs. There is just a slippery slope between monitoring for diagnostics and abuse that incremental business value very quickly slides into tying team-specific view counters and a favorite team to my user identifier.

Yes there are other solutions that don't require Google to know who your favourite team is, but needing to have your device run a specific background process to constantly poll a remote server for each of the things you're interested in is a pretty clunky and resource-inefficient way of solving it vs subscribed push notifications from the server.

And Google doesn't necessarily know that you like $team. It knows that the email address you used to sign up to the alerts has requested updates on it. Is that really particularly scary?

I'd like Google to store only at most N bits of information about me. That way, when I type "python" in the search bar, Google knows I'm a programmer, and am not interested in results about snakes. However, google does not know what television show I watched last night, because that doesn't fit in the N bits that I'm allowing Google to store about me.

Viewed differently, the value of N specifies how uniquely identifiable the user is.

On-premise vs. Cloud?

I know it's been beaten to death. The cloud has been winning over last decade, we'll just have to wait for the next wave to come and bring back the intelligence next to the user.

Reply on news.ycombinator.com