Scamdroid : scambusting ai and decentralized web app aka what savedroid should of done


#1

Help me?

You guys are smarter than me.

I’ve decided to put wheels on my rant and combine my deep dives for the rest of the year
deep dives= full stack webdev, ICO analysis, blockchain (as in protocol level), and AI. (thanks to Mr Fishy who gave me a quick window into trading bots on the forum that has my hacker brain inspired). I’m not a programmer but I can program and my work in crypto warrants the investment of time given the value in the learning.

So smart people which means (not @peter) where would you start? I wanna stick with Node and JS as much as possible. Frameworks?

The thing most interesting to me right now is how would you teach a robot to spot a scam ICO and rank the likelihood of skullduggery.

Ok that’s all interested in your thoughts (again not @peter)


#2

You could try scraping this site: https://icowatchlist.com/

The rest is not going to be easy.

Good luck.

Mike Fishy


#3

Yeah for sure. I’m just going to mostly talk to myself in this thread. Feel free to interupt me and tell me I’m an idiot. Again not an engineer, programmer, etc. so I get to be stupid.

Simply, the goal is to automate the process of evaluating an ICO like an anti-virus evaluates a file for malicious code. But scanning through files on a computer against a database of malicious code is pretty easy when compared to the ambition of automating the process of evaluating an ICO for scam indicators. Forget risk we’re just looking for scams.

Scraping a list of ICO’s future, current, and past is a good start. In theory my little robo cop should collect data and monitor every ICO indexing site it can find (not just in English) that’s a bunch of work right there ;-). Not very cool but useful

That gives me a base index and simple dataset with coin names, logos, urls, maybe whitepapers, etc. which will come in handy when I present these trolls to the public or at least to my human eyes…I like graphics.

but I want my little buddy to do way cooler stuff so I need a deeper dataset for each ICO

“Though not at the forefront of the AI hype train, the unsung hero of the AI revolution is data — lots and lots of labeled and annotated data, curated with the elbow grease of great research groups and companies who recognize that the democratization of data is a necessary step towards accelerating AI.”

The next thing I want to look at is all of the people attached to the project. Now that’s tricky…so I call my friends at LinkedIn, Facebook, Twitter, etc. How do I evaluate a human with a robot? Maybe I don’t…this feels like a job for a human or better yet a crowd of humans who are incentivized to help me collect the data through some kind of crypto fueled bounty program :wink: For instance @peter who is this guy? what has he done in the past? what information is publicly available about him? where did he got to school? what did he study? how can I verify what his academics or previous work history? How long has that information been on the web (did it just suddenly appear or has it been there for years?) are there news articles about him?
what are the titles of the news articles? when were they written? what is the quality of the publication who wrote them? what information about peter isn’t public? can I get legal access to it through a background search service API? how many times was peter sued? How many times did peter sue someone? Was peter ever arrested? did he ever commit a financial crime? Holy crap this is getting complicated…I give up :wink: at least for now…


#4

Prolly better

Of course there’s this TOU:
1.1.3 Use any robot, spider, crawler, scraper, or other automated means or interface not provided by us to access our site or to extract data;


#5

I actually had this thought myself. My first evaluation would be the site itself. Is it just a template with broken links on free WP? I see so many and think who would fall for this shit it is such a bad site! Anybody trying to raise a ton of money should at least invest in a better site.


#6

So a Hallmark of scams is the paid shill right? So how would you monitor and measure promotional channels for things like sudden spikes in promotion, proliferation mentions etc What would those channels be? How would you monitor them? @savedroid for example they ran such a successful ICO I heard they are going to consult now…maybe I’ll check out how they ran their promotion as a baseline. I’ve heard their marketing team is so cutting edge it’s bloody. Then I’ll move on and analyze Pincoin and maybe Centra.

Obviously legit ICOs do the same stuff but what if there’s a tell in their promotion. Like the YouTuber they do a deal with or the website they pay to run an ad etc

How do you teach a robot to evaluate a scammer shill campaign?