Live data from Hacker News

Statistics Professor Just Banned by Google: Here Is His Story

investingchannel.com

131–140 of 211 posts

Re: Statistics Professor Just Banned by Google: Here Is His Story

#131
post #97

Earlier quoted context omitted.

Its not the mail server, its deliverability. If you can set up a $5/month server to send email so can every spammer ever and the big mail providers know this. Mail sent like this won't be delivered anywhere with even the most basic spam filtering which includes blacklists of known spam-source IP address blocks like Digital Ocean/AWS/Heroku.

I have a personal mail server that collects my mail, but use my "default" Comcast MX to send. I'm sure metric craptons of email flow through that, but a lot of ham does too, so I've never been blacklisted. I'd suggest that rather than the goal being to instantly leave this service or that the instant that they offend you, that you envision your policy as one of being ready to leave if you ever need to. The biggest ke…

What prevents your domain name registrar from shutting you down arbitrarily?

Also, I'm not sure what you're referring to as Comcast MX, but if you mean a Comcast provided server, that's a huge privacy hole. If you mean your own server on a Comcast line, then your IP is leaked to everyone you email, which is also a privacy hole since it ties your IP directly to your identity.

Re: Statistics Professor Just Banned by Google: Here Is His Story

#132

It seems that his ban may be political in nature according to his tweet. https://twitter.com/salilstatistics/status/89891222812829696... Salil Mehta: "Shameful: if you show probability work like Hillary having lower election odds, then this is new definition of hate speech. @JulianAssange"

Seems Is that tweet in violation of Google's Terms and Conditions? edit: formatting

Twitter is not owned by Google. Google's ToS cannot interfere with his activity on a property they do not own. That would be tortious interference of contract.

Re: Statistics Professor Just Banned by Google: Here Is His Story

#134
Google Takeout

https://support.google.com/accounts/answer/3024190?hl=en

https://en.wikipedia.org/wiki/Google_Takeout

Problems though: if your account is already disabled, you can't use this as far as I can tell; and there's no diff option, it's a monolithic backup each time. But it might provide a strategy to limit data loss resulting from having an account disabled for whatever reason.

Re: Statistics Professor Just Banned by Google: Here Is His Story

#135

Earlier quoted context omitted.

Google's AI jumbling up my search queries and insisting on answering entirely different questions from the one I asked has finally pushed me to duckduckgo (which is greatly improved and not yet as botty.) So I bet you're right thinking it's a demented bot. Such bots (and especially, "forests" of bots) tend to punish anything that's sufficiently eccentric. This could be countered by a separate forest of bots that look…

Fwiw. Duckduckgo is using Bing under the hood I believe.

That's a bit of an oversimplification. From the FAQ page:

> When people search, we believe they're really looking for answers, as opposed to just links. For many categories of searches (restaurants, game of thrones, code documentation, etc.), there is usually a specialized search engine (Yelp), content site or other source that does a better job at actually answering the searches than a general search engine does with just links. Our long-term goal is to get you Instant Answers from these best sources.

> To do that, we've developed an open source Instant Answer platform called DuckDuckHack, which allows anyone to propose and develop Instant Answers for DuckDuckGo. On the backend, we have a classification and relevancy engine that selects the best sources for a particular search and displays answers to you in real time, be it for recipes, products, images, and hundreds of other categories.

> In fact, DuckDuckGo gets its results from over four hundred sources. These include hundreds of vertical sources delivering niche Instant Answers, DuckDuckBot (our crawler) and crowd-sourced sites (like Wikipedia, stored in our answer indexes). We also of course have more traditional links in the search results, which we source from Bing, Yahoo, and Yandex.

https://duck.co/help/results/sources

EDIT: changed link to more elaborate FAQ one.

Re: Statistics Professor Just Banned by Google: Here Is His Story

#136
post #69

Earlier quoted context omitted.

Do you find you have problems delivering to other mail servers? That has always been the trick. Running the mail server is easy. Having your email show up in other servers is "hard". Right now, with so many users using GMAIL, using gmail is a huge boon to not being marked spam (Google is way less likely to market a @gmail.com email as spam compared to say a @aol.com email).

I haven't had any issues but I've been hosting my own mail server since 2000 or so. I've been blacklisted by spamhaus? before years ago but I submitted a request to remove and they did. Google has a page to help with this though. https://support.google.com/mail/answer/81126?hl=en&vid=0-107... Also, I personally host at home and have Comcast business internet. I have a 5 IP block of which one is a mail server. I have…

Your IP is leaked to everyone you email (and if you receive mail, to everyone who emails you), which is a privacy hole since it ties your IP directly to your identity. I assume you don't mind, or did you not realize this?

Re: Statistics Professor Just Banned by Google: Here Is His Story

#137

Earlier quoted context omitted.

I'd like to add another voice to this: I host my own email myself out of my house, and I've never really had any problems with it either. AFAIK I've never been marked as spam, even after I've sent a couple of mis-configured emails that should have probably been marked as spam.

Have you ever done mailing list work? Like run a mailing list with a few tens of thousand users?

Nope. I can't say that I have. That definitely might be something that can get you flagged.

Re: Statistics Professor Just Banned by Google: Here Is His Story

#138
post #85

I wonder how much I would lose if my Gmail was deleted... Everything I have is behind that address. What would be my best bet to move to if I'm too much of a bad to host my own email?

I do two things: 1) I use an address on my own domain (hosted on Google Apps) as my primary email address. If Google pulls the plug on me I can point my MX records elsewhere, and won't have to worry about changing my email at countless places. 2) I use offlineimap to back up the contents of my GMail account at home. If I had to, I could load that backup into a new provider's service (or my own). I still have email fr…

Of course 1) only works until your registrar decides they don't like the content of your website.

But you can back up virtually all data hosted with Google through Takeout.

https://takeout.google.com/settings/takeout

Re: Statistics Professor Just Banned by Google: Here Is His Story

#139

Earlier quoted context omitted.

Fwiw. Duckduckgo is using Bing under the hood I believe.

That's a bit of an oversimplification. From the FAQ page: > When people search, we believe they're really looking for answers, as opposed to just links. For many categories of searches (restaurants, game of thrones, code documentation, etc.), there is usually a specialized search engine (Yelp), content site or other source that does a better job at actually answering the searches than a general search engine does wit…

This reminds me: google owns duck.com

Don't mistype duck.co

Re: Statistics Professor Just Banned by Google: Here Is His Story

#140
post #47
post #2

No idea what all was on his blog and other google accounts, but according to the article they were used for his work at the university. Also some credible sites like the NYT linking into the disabled blog: https://www.nytimes.com/2017/08/19/business/the-stock-market... Google may have had some good reasons or it may have just been an example of ML discriminating against a name like Salil. Either way, it's very distur…

I strongly recommend everyone to look into hosting your own mail server. DigitalOcean will rent you a Linux VM for $5 a month or something crazy like that. You can figure out how to install postfix/dovecot/squirrelmail it in a few hours. If nothing else, it keeps critical functionality decentralized, and you learn something and $5 a month for career insurance is pretty damn cheap.

How good are open source spam filtering tools? Besides not having to maintain my own infrastructure, spam filtering is probably the biggest value add for Gmail, or so it was when I got my account 10 years ago.
Post reply on HN