Hi folks, I'm sure you have seen @Pizzagatebot commenting by now. I started the thread here last week - https://voat.co/v/pizzagate/1676126
Pizzagatebot is off probation and officially running on automated schedule on its own linode server.
It took a bit longer than I thought it would take since I didn't get a voat API and decided against using googles API, added some features to my host server like VPN, proxying and other tricks to try and avoid being tracked. The advantage is that all the code is my own so I know its safe and I can change any of it, vs taking someones off-the-shelf code from github, checking and learning their code and trying to modify/make it work for my needs.
Due to some peoples comments, I would like to reiterate that the bot is there to help link information together, not shame you for not posting those things in the first place. I edited my post with a new link and it ended up commenting in my post here - https://voat.co/v/pizzagate/1684578
If you see it break or linking something ridiculous, just PM the bot with the permalink, I will check it from time to time.
If you folks have any other ideas for things that should be automated to help the community, let me know, can't hurt to mention it. For example, I have a few new test functions finally working that I will roll into the main code later this week such as:
- reverse image search to find posts with the same/similar image posted
- searching against deleted submissions to help some self moderate and alert other readers to this fact
- archive.is link resolve to search for original link
I chose against listing the date and upvoat/downvoat of the posts since I think it may discourage someone from finding something that may have otherwise continued to be overlooked and I don't really see it linking to anything that I would consider solid disinformation right now.
I can block it from searching for links to certain sites, I thought about it for undergroundnewsreport.com since it seems to be disinfo/COINTELPRO related. Any comments on that would be great.
Next up to bat is trying to figure out how to use different data analytics libraries for processing email/facebook/twitter data scrapes and see if there is anything interesting from automated correlation there. We need to use their own technology against them.
Thanks to everyone for keeping this community alive and I hope you all stay safe, aware, patient and vigilant. Cheers!
view the rest of the comments →
SaneGoatiSwear ago
fucking cool. 2 things: any chance moving away from google (maybe towards searx.me - which can proxie through to google anyway), and 2. thoughts on other types of helpful bots you're thinking of? (ahem, shills... so many shills)
throwaway89209834 ago
Yes, I know about google and their controversies, but google still peforms the best, seems to crawl new pages the fastest and it's just easy to work with. I actually tested with Yandex since you can force it to do exact match always, but Yandex didn't return some same results in compared searches.
I like that searx.me is open source, I'll put it on the list to test later.
I should of put in the first line of the post 'i will assume you are all thankful, only comment with ideas or thoughts on this' because I am looking for ideas and wanted more of a brainstorming session.
Searching against deleted posts and commenting with some suggestion to delete will help with some of that I think. Or maybe checking if its a discuss post but has no links and reminding they need to have references.
Not sure what you are thinking, but it can't be anything to publicly label or suggest a shill because someone could be falsely labeled. It has to stick to a scope of suggestions that can't be taken as accusations.
SaneGoatiSwear ago
ah, then, specifically: disinformation tactics: https://cryptome.org/2012/07/gent-forum-spies.htm and fallacies: https://en.wikipedia.org/wiki/List_of_fallacies
i appreciate your full response above, either way, and that you'll test searx; i understand google's ease. good thing it's not called adam's apple! fuck organized religion.
anyway, yeah, you rock. bot rocks. keep kickin ass behind 77 proxies because every time some good coder pops up here, they are soon replaced with some angry anti-speech version. it's scary. sli.mg devs are still missing 7+ months from the internet, jmtullock was replaced with angry version. the same day the slimg devs disappeard so did atko and puttitout, and they came back 1 and 2 months later angry distant aloof versions of themselves as well. watch your ass is what i'm saying.
throwaway89209834 ago
Woof, if someone comes up with a way to detect those type of fallacies in a blob of text, they will make millions. I have no idea what is cutting edge in this area right now, but I haven't heard of anything like that anywhere yet. I'm not a data scientist, just have some coding experience and trying out different things, like extracting proper nouns, unique phrases, keywords and seeing how relevant it is and how it could be applied or analyzed with other data.
Thanks for the warning too.
SaneGoatiSwear ago
"neural nets"
throwaway89209834 ago
Yeah that came up in research into data analytics, but have you heard of anyone actually successfully doing that? There has to be hundreds if not thousands of applications. If someone got it to be reliable you'd think you'd see it at least being patented if not already being licensed.
samhara ago
Machine never passed Turing Test yet.. It's hype.
throwaway89209834 ago
I agree. Keep in mind a lot of high tech stuff in the private sector came from and still comes from public sector. During/before WW2, the saying was that the gov was 50 years ahead of current consumer technology and now, with the rise of corporations, the crazy high tech stuff is in the corporate R&D labs. Which is of course favorable due to corporate personhood and less oversight and we know the CIA is well vested into various tech companies..