HNHacker News
TopNewBestAskShowJobs

kristintynski

879 karma · joined October 24, 2013

submissionscomments
kristintynski··on Cryptocurrency scams cost owners $7.7B in 2021, driven by DeFi-based “rug pulls”
Somehow managed to avoid like 5 of these memecoin rugpulls last year. Memecoins are a clown casino, but the memetic dynamics fascinate me.

The right core meme can be a catalyst for compounding user generated content growth. Look for memecoins with community growth metrics that look exponential, make sure the tokenomics are safe, and hold on for dear life.

For me, DogeBonk and perhaps one or two other memecoins have the community growth and memetic qualities to be exponential. So long as there’s no rug.

kristintynski··on How to Start a Movement – Strange Experiences in Marketing
Continued...

2. Dogebonk I briefly got into crypto in 2018, and left shortly after, disillusioned with how unpredictable and volatile the whole space seemed. In the last few months, however, I’ve found myself back on the hype train, trying once again to try and grasp how the blockchain and Web3 will change the world. I’d watched with fascination this year as NFTs and Memecoins started taking over, even though from afar it seemed like a total shitshow. That assesment wasnt far off. What I was wrong about however, was that there were few to no opportunities there, and what really shook me was that once I got involved, I realized they were doing marketing on a level I’ve not encountered in my 15 years as an agency owner.

I’ll preface the rest of this story by saying what I’m about to tell you is among the riskiest bets you could possibly make, and it is something I treat as a moonshot, and not as something guaranteed. So please do not take this as an endorsement, instead look for the commonalities in this example and the first.

Still, what I’m going to tell you about is the most impressive and unusual thing I have witnessed as a marketer.

About two weeks ago I stumbled upon a Reddit post for a new memecoin that I assumed was a scam, but it got me thinking. I’ve been casually investing in crypto for a few years, but always stuck to the primary players. I’d watched the memecoin and NFT frenzy with amusement, but honestly thought it was ridiculous. And truth is, it is more than ridiculous, it is the wild fucking west with everyone high on crack.

After witnessing the rise of Shiba Inu after Dogecoin, I felt like I needed to understand how exactly millions of people were deciding to make these seemingly insane investments. Is it all just pure speculation? Well, sorta, but there is a lot more to what is driving it than just the promise of money, and the community dynamics and the content they create are what matter here.

Back to the suspect new memecoin. What I stumbled upon that day was a memecoin called Dogebonk on the frontpage of Reddit, where people were griping about a Billboard they had purchased in Times Square. Having known about the power of a single viral post to drive the creation of communities with common goals from my experience with PDAP, I wanted to see if the same process was happening here as well.

I joined their telegram, and was immediately hit with a rapidly scrolling chat of memes. With a few thousand members, 1/3 of them were active in the telegram at any given time. I’ve been working in digital marketing for 15 years, and have very rarely seen such levels of community engagement. This was all within a day or so of a single viral post on Reddit.

It hit me, memecoins are simply the monetization of memes. Take an existing meme, and a community that already exists will pump out those memes quadrupole time if they think their memeing will help grow the coin. ( which it will if enough people in the community do the same).

What you get is a flywheel that is essentially a viral marketing machine, with individual community members promoting the coin in such variety and scale, any marketing company would be paying millions per day for an equivalent reach. This attracts more fans of the meme, more investment, and more members to continue to grow the community and thus the value of the coin.

When you pair a quality meme, with the power to monetize that meme as a coin, the limiting factors are the quality of the meme itself and the skill of the marketing of the community.

At first glance, dogebonk sounds as dumb or random as any other memecoin, but as I’ve found, there are some critical differences.

The first is the meme quality. What makes dogebonk as a meme intriguing is its extensibility. Because it is an “active meme,” a dog bonking another dog, that action can be capitalized on in a way that other memes can’t. This subtle difference is striking when you look at doge or shiba vs dogebonk, the volume of memeing is vastly more common in dogebonk. The more frequently and easily a meme can be iterated and riffed on, the more you will see viral memes emerge. These memes are a fuel for the community, creating cohesion as well as fun. Each new hilarious meme has a dual purpose of solidifying in-group identity, and creating a vehicle for external recruitment. It’s an internal engine that keeps even loosely tied communities together. This was one thing PDAP lacked, and in the weeks that followed my joining of the Telegram group, I realized how critical that would be to maintaining a large percentage of initial true believers.

What I witnessed after joining dogebonk, was in three days they had gotten a Times Square digital billboard, and then frontpaged a Reddit post about it. Then I saw them take over the front page of a bunch of crypto subreddits, meme subreddits, and /biz on 4chan. The collaborative power was mind boggling. With the right meme, a brilliant grassroots community “bonksquad” of meme making and expert marketing has emerged and grown massively, all while utilizing memes to keep high morale, engage existing coin holders, and provide fodder for extending visibility..

The key both examples here is that organic community growth happens when you have two things in common.

1. Something people can get behind and work together on that has a concrete and tangible benefit to them or the things they find important.

2. A highly visible initial post that directs viewers to a community where they can immediately participate and engage.

These two examples are just two that i’ve witnessed recently, and it informed my opinion significantly on the power of a small spark, understanding that there is a real path from complete obscurity to massive community power, and that path isn’t as hard to travel as you might assume.

kristintynski··on Ask HN: Has Google search become quantitatively worse?
GPT3 got your back:

warmest winter coat --hobbyist-only

Top 10: 1. Canada Goose Parka 2. Patagonia Down Jacket 3. Marmot Precip Jacket 4. Columbia Winter Jacket 5. North Face Thermoball Jacket 6. The North Face Nuptse Jacket 7. Rab Neutrino Endurance Jacket 8. Mountain Hardwear Ghost Whisperer Jacket 9. Black Diamond Fineline Hoodie 10. Outdoor Research Cathode Hoody

kristintynski··on Crypto’s hottest game is facing an economic maelstrom
fascinated by the memetic properties of dogebonk. viral loops galore
kristintynski··on Mastering Web Scraping in Python: Crawling from Scratch
Shameless request for scraping enthusiasts at: https://www.pdap.io, an open source Police Data Accessibility project started on HN and Reddit. Our goal is scraping and collating all county level public records, giving us a dataset to enable "Policing the Police"
kristintynski··on GAN Theft Auto [video]
No one cares how you feel, come with proof before accusations. Otherwise you are just a troll
kristintynski··on GAN Theft Auto [video]
Accusations of scamming are serious. What evidence do you have? None as far as I can see. This is wrong and should be remedied.
kristintynski··on Show HN: Policing the Police with Scraped Data – Started on Reddit/HN
Predictive policing is rampant and growing. Why not try and do predictive policing on the Police themselves?

The road block isnt data availability, it is data accessibility.

PDAP.io started on Reddit and Hackernews, and as of last month we've accomplished a lot:

1. Became a legal entity 2. Pro-bono legal representation by Arnold & Porter 3. Leadership team, roles, processes stood up 4. Launched our first paid data bounty with Dolthub 5. 37 Pull requests merged into our scrapers repository (dozens of scrapers written) 6. 67 datasets scraped so far 7. Nearly complete dataset of all known American police records links (from the Dolthub bounty).

What we need now are more volunteers, specifically people who can write scrapers. Things are moving quickly, but with just a dozen or so people writing scrapers, we could accellerate this substantially with more help.

So, HN, if you have wanted to do something tangible to improve Police accountability, this is a concrete way to do it.

kristintynski··on New York makes complaint records of 83,000 police officers available to public
We can police the police by tracking them with data. Complaint records are amazing to have, but we can also find dirty/racist cops with existing data from county public records. It just needs to be scraped and collected. Redditors and folks at HN started this project, and we are making progress, but it's a big task. With enough volunteers it will be possible to create a database and scraping infrastructure that will enable citizens to truly understand how individual police officers and departments are behaving. Context: https://www.reddit.com/r/privacy/comments/m59o2g/i_think_i_a...
kristintynski··on Ask HN: What (side-)project are you working on?
HTTPS://PDAP.IO a nonprofit with the goal of scraping and making accessible county level police data. With this data aggregated and made available, policing the police will be more possible.
kristintynski··on Policing the Police by scraping court data – *An Update
These are the fields we are currently scraping, but we ask devs to please just comment out lines for scraping officer names and badge numbers. This was simply out of caution as we finalize the nonprofit application.

The intention is to have datasets of individual officer level behavior, especially around stop/arrest level data and arrestee level demographic data.

kristintynski··on Policing the Police by scraping court data – *An Update
All types of infractions. https://github.com/Police-Data-Accessibility-Project/Police-...
kristintynski··on Jacquard by Google
The app for soccer players seemed compelling as a use case for competitive teams. Perhaps this is already being done by other means, but it seems like you could pretty simply track detailed performance data at the individual player level and whole team level. I can see it being big in this space.

Also, maybe health applications? Vital monitoring? Posture tracking? who knows. To me, any new platform like this opens up at least some interesting opportunities for niche use cases.

kristintynski··on How to buy gifts that people actually want
I made this with the GPT3 api. It’s a gift recommendation app. HTTPS://GiftGenius.AI
kristintynski··on The Police Data Accessibility Project
yup, this is the same group/movement
kristintynski··on U.S. Police Shootings: Breaking Down the Data by Race
Now is the time for us to start doing the hard work to make police accountability happen. Getting public records data accessible is a way to make good progress toward this.

450+ of us are working on a new project called the Police Data Accessibility project, aimed at making local court record data (which contains cop/department level data) accessible outside of clunky county court records websites. More info here: https://www.reddit.com/r/privacy/comments/gr11aw/i_think_i_a...

kristintynski··on Show HN: I scraped local court records to find dirty cops
I'm much more interested in making this data open and complete and potentially close to live updating as possible.

I agree that it will take more depth to get at true patterns in the data, and there are many important considerations for data collection, cleaning, and analysis.

I hope to find some others who would like to work with me on this.

kristintynski··on Show HN: I scraped local court records to find dirty cops
I'm beginning to think it might be the start to a more just society. Is that thinking this is bigger than it is? The implications just seem big to me. I don't think we are going to get privacy back, but if we can at least have transparency into things like policing, we stand a much better chance at holding our government accountable.
kristintynski··on Show HN: I scraped local court records to find dirty cops
This is SO useful. Thank you so much for the insights. I'm not 100% certain, but I think finding traffic stop data is pretty rare for most counties.

If citation data is all we have, by itself in its raw form is still important considering how locked away this data is currently.

I think there are a lot of smart people like yourself that would do in-depth analysis of even this messy data to pull out useful insights.

I'd be eager to get people like you involved if you had an interest. you can email me at:

Kristin - at - Frac.tl

That goes for anyone who has an interest in helping. :)

kristintynski··on Show HN: I scraped local court records to find dirty cops
I agree with you, and that's why the officer names are redacted. Things we might at first glance think are because of biased/racist/sexist cops might have good explanations. That said, visualizing this data does surface outliers really well, and each outlier should be further investigated IMO.

It all starts with getting all this data into one place, scraped from the thousands of county court records sites. It's a big task, but this data is priceless.

kristintynski··on Show HN: I scraped local court records to find dirty cops
It's a matter of public record, though I guess they could gripe about any data viz methodologies... still, we need this type of oversight.
kristintynski··on Show HN: I scraped local court records to find dirty cops
I plan do do this for many more counties. I'm sure it isn't a coincidence how difficult many are to scrape. This data often feels purposefully obfuscated. It only seems right to me that everyone should see local data aggregated like this. If police intent do use data to track the population, we should be able to track them as well.
kristintynski··on Show HN: Scraping Google's Covid Mobility Reports – US South in Trouble
Strangely, Google never released maps of this important data, which shows which counties are heeding "stay at home orders," So, I scraped the data from the PDF's they released, and mapped it. I think this is one of the first geographic views of highly detailed information on behavioral changes due to Covid-19.

Sadly, it seems red states (in particular the US south) could really be in trouble.

kristintynski··on [dead]
The article doesnt claim he does or doesn't. It just looks at the data. The conclusion explains there isnt any conclusive evidence either way. Did you read it?

While it may seem “obvious” to some of us that Trump isnt the man he used to be in terms of cognitive or speaking ability, sussing out proof of this decline is harder than one might imagine. Although there are some clues, they are far from conclusive. By comparing Trump with Sanders, we see similarities in changes over time, perhaps with Trump’s changes being mildly more pronounced. Whether this means both Trump and Sanders have declining cognitive abilities is an open question. For now, it appears there isn’t enough of a change over time to find anything resembling conclusive evidence for Trump having dementia, Alzheimer’s or other cognitive decline.

kristintynski··on I used AI to show Trumpers are more easily fooled by AI fake news
It has its role in quick and dirty preliminary analysis and experiments. Studies have shown it is actually pretty damn good. http://web.mit.edu/bentley/www/papers/chi-case-surveys.pdf
kristintynski··on Show HN: RoboTrump – I made a Trump text generator that is fully convincing
that is the next logical step in this fake news dystopia isnt it. Seems like someone could relatively easily set up a pipeline to generate text, turn text into Trump's voice, then apply to Deepfake video, churn out millions of videos... profit. kill democracy. etc.
kristintynski··on Show HN: RoboTrump – I made a Trump text generator that is fully convincing
As long as I stuck to prompts related to politics and things Trump has talked about, it was scary good. I'd imagine fine-tuning on a much larger Trump dataset would continue to improve it.
kristintynski··on Show HN: RoboTrump – I made a Trump text generator that is fully convincing
I didn't really cherry pick, most everything it generates is of this quality.
kristintynski··on Show HN: RoboTrump – I made a Trump text generator that is fully convincing
I can upload the trained model if you are curious, but it was straightforward to do. I just used the Salesforce CTRL model and fine-tuned it on a large corpus of Trump speeches.
kristintynski··on RoboTrump – A fully convincing Trump AI text generator
It could be easily trained with additional fine-tuning to have whatever bias you want him to have that is different from his current biases. What makes it so scary is the volume of realistic text that can be generated. It could be used to make Twitter/social media bots, or any volume of fake news articles, or combined with deepfakes and AI voice cloning to make unlimited volumes of fake video/audio of him saying things, in his style, about any subject with whatever angle/message you wanted.
← PreviousPage 3 of 4Next →