Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Computer, Respond to This Email

googleresearch.blogspot.com

Post reply on HN
Print view View original

161–170 of 202 posts

Re: Computer, Respond to This Email

#161

Post by keshav57 »

> The solution was provided by Sujith Ravi, whose team developed a great machine learning system for mapping natural language responses to semantic intents. This was instrumental in several phases of the project, and was critical to solving the "response diversity problem": by knowing how semantically similar two responses are, we can suggest responses that are different not only in wording, but in their underlying meaning.

Does anyone know where I can find out more about this? A paper or something maybe?

Re: Computer, Respond to This Email

#162

Post by zhanwei »

next step, check your calendar for scheduling conflict and suggest the right response, i.e., "i'll be there"/"i can't make it".

further extension, pull out quick answers from other emails, i.e., "what's the sales number this month?" "100M" from another email

Re: Computer, Respond to This Email

#163

Post by WWKong »

The real potential here is eliminating human interaction. On the sending end machine should auto compose, auto target a mailing list, send it out. On receiving end machine should scan incoming email, compose a response, send it out. That would cover 95% of current email traffic.

Re: Computer, Respond to This Email

#164

Post by dikaiosune »

TeMPOraL wrote: post #154
dikaiosune wrote: post #25

Earlier quoted context omitted.

Because I'm lazy and switching would be time consuming. Each new privacy-related "innovation" motivates me a little more. You're right that the practical effect of this new service is negligible on a technical level, but it still reminds me that I should find an alternative I find acceptable.

So basically, because pretty much every useful data-and-automation-based thing - especially on the social, not individual level - can be considered an erosion of privacy, should we just stop progress? That's the sentiment I feel radiating from yours and others' complaints about privacy here.

Certainly not. I quite enjoy the features I get from having my data run through a lot of systems. However, you don't need to be a Luddite to recognize that sometimes some users of data (like sometimes Google) aren't as responsible or considerate of the privacy implications of their actions as they should be.

When their primary financial incentive is user engagement to sell ads, I have a hard time believing that the latter doesn't influence the use of my data as much as the former.

I still use a variety of Google products all the time, but when they're training machines to understand the meaning of my emails, I begin to think perhaps that's beyond my comfort zone. I think that machine learning has tremendous things to offer society, but I'm not sure I'm super excited about a company using smartish machines to pull meaning and context from my emails so that we can save some percentage of users 10 seconds when they reply to something. I get that in most cases this isn't ever going to have a practical effect, but I would be a lot more comfortable if our society's data stewards were less blase about that side of things.

Re: Computer, Respond to This Email

#165

Post by TeMPOraL »

dikaiosune wrote: post #164
TeMPOraL wrote: post #154

Earlier quoted context omitted.

So basically, because pretty much every useful data-and-automation-based thing - especially on the social, not individual level - can be considered an erosion of privacy, should we just stop progress? That's the sentiment I feel radiating from yours and others' complaints about privacy here.

Certainly not. I quite enjoy the features I get from having my data run through a lot of systems. However, you don't need to be a Luddite to recognize that sometimes some users of data (like sometimes Google) aren't as responsible or considerate of the privacy implications of their actions as they should be. When their primary financial incentive is user engagement to sell ads, I have a hard time believing that the l…

I see your point better now, thanks for clarifying.

> When their primary financial incentive is user engagement to sell ads, I have a hard time believing that the latter doesn't influence the use of my data as much as the former.

I don't agree with it in case of Google. I may be mistaken, but my impression was that they sort of separate their products into two groups - the ad-related are earning the money, and then the money is spent on funding something else (like GMail), with little direct connection between the two groups. That is, products like GMail certainly help Google earn more money on ads, but are not themselves optimized for ad-related purposes.

> I think that machine learning has tremendous things to offer society, but I'm not sure I'm super excited about a company using smartish machines to pull meaning and context from my emails so that we can save some percentage of users 10 seconds when they reply to something.

I think you're seriously underestimating the potential for productivity gains here. Mobile use case is perfect for replying to e-mails, but mobile experience totally sucks. This service, if it works as advertised, will be probably saving not 10 seconds, but something like one minute per e-mail. That + friction reduction have enabling properties that make people do things they didn't before (e.g. I often don't reply to e-mails on a phone only because it's too slow, opting to browse Facebook instead), and aggregated that can help liberate a lot of time.

Re: Computer, Respond to This Email

#166

Post by dredmorbius »

Maxious wrote: post #126
mmanfrin wrote: post #93

Earlier quoted context omitted.

I had a similar thought with Apple -- El Cap has a feature where if you shake your mouse around, the cursor will momentarily grow in size so you can find where it is. That does not feel like a top-down design idea, but a bottom-up feature designed by an engineer who was sick of those moments where he/she had lost the cursor.

I seem to recall the Microsoft Mouse drivers for Windows 95 had a key shortcut to make lines radiate out of your cursor. Very useful on the low refresh rate LCDs where that cursor would appear like a submarine on the other side of the screen.

I liked the "shroom mode" mouse option in one of the hack packs.

Trails, man....

Re: Computer, Respond to This Email

#167

Post by dingo_bat »

supersan wrote: post #52

I'm often amazed how Google still gives it staff this unique hacker-like approach to their million dollar projects. For example, the stuff in Google labs (gmail) has had some silly things like don't hit send when you're drunk but sometimes very useful features which may be considered unorthodox like "Gmail telling you when it thinks you wanted to attach a file but forgot". It's something I wouldn't generally associat…

"Gmail telling you when it thinks you wanted to attach a file but forgot" is hardly unorthodox considering Outlook has had it for years. That said, the other experiments do seem very informal and fun.

Re: Computer, Respond to This Email

#168

Post by techdragon »

rylee wrote: post #127
Maxious wrote: post #126

Earlier quoted context omitted.

I seem to recall the Microsoft Mouse drivers for Windows 95 had a key shortcut to make lines radiate out of your cursor. Very useful on the low refresh rate LCDs where that cursor would appear like a submarine on the other side of the screen.

You can still do this (at least in Windows 7, the last time I checked.) Just look in the 'Mouse' section of your control panel, and select the one to make radiate circles around your mouse :)

I can confirm it still works in Windows 8/8.1

Re: Computer, Respond to This Email

#169

Post by rahimnathwani »

mmanfrin wrote: post #93
supersan wrote: post #52

I'm often amazed how Google still gives it staff this unique hacker-like approach to their million dollar projects. For example, the stuff in Google labs (gmail) has had some silly things like don't hit send when you're drunk but sometimes very useful features which may be considered unorthodox like "Gmail telling you when it thinks you wanted to attach a file but forgot". It's something I wouldn't generally associat…

I had a similar thought with Apple -- El Cap has a feature where if you shake your mouse around, the cursor will momentarily grow in size so you can find where it is. That does not feel like a top-down design idea, but a bottom-up feature designed by an engineer who was sick of those moments where he/she had lost the cursor.

This feature is awesome, and super-discoverable. I didn't need to change my behaviour to discover it, or to use it. I just wiggle my finger like I used to, and it just works.

Re: Computer, Respond to This Email

#170

Post by microtonal »

TeMPOraL wrote: post #154
dikaiosune wrote: post #25

Earlier quoted context omitted.

Because I'm lazy and switching would be time consuming. Each new privacy-related "innovation" motivates me a little more. You're right that the practical effect of this new service is negligible on a technical level, but it still reminds me that I should find an alternative I find acceptable.

So basically, because pretty much every useful data-and-automation-based thing - especially on the social, not individual level - can be considered an erosion of privacy, should we just stop progress? That's the sentiment I feel radiating from yours and others' complaints about privacy here.

should we just stop progress?

I would love if companies would explore on-device processing more. While it might not work in this case, it should work fine for e.g. data extraction for flight tracking, hotel bookings, package tracking, etc.

Post reply on HN