I mentioned it briefly in the blog post, but this is exactly what I'm working on! Essentially, a spiritual successor to Fakespot that combines LLM analysis, more traditional ML techniques, and rule-based heuristics to detect fake reviews. I'll likely go the "subscription with generous free trial" route, to avoid meeting the same fate as Fakespot.
I'm actively working on a prototype and have a landing page at https://www.truestar.pro if you want to get notified about when I launch.
Please help me understand why a subscription to your service should be a valuable addition to my monthly spend.
I buy extensively from Amazon across a number of product categories. My order history shows purchases as far back as 2005 (though I cannot be sure given I remember buying things in 1998 while in college, probably on a different account). During the intervening 20 years I can count on one hand the number of products I ordered which weren’t legitimate, matched my—admittedly moderate expectations for any commercial product—or included overhyped reviews.
I’d be interested in a service like yours if I could understand how the cost would cover itself in benefits.
Attribution revenue is what I would consider the gold standard for these types of services.
It makes sense on paper, if the service helps confirm legitimate reviews for you and convinces you to purchase said product, they should receive attribution revenue for helping generate the purchase.
The reality is much much messier though, because often times the people who award attribution revenue have a conflict of interest against any service that could even potentially expose bad practices happening on their marketplace.
I once worked for a popular deal site that developed a price tracking extension, a certain marketplace threatened to completely ban us from attribution revenue and we had to kill the extension over night despite our users loving it.
That does not help me understand why I should pay for this service. I personally have zero concern for why this service should make money beyond operating costs. “Why should I should pay for it” is the question I asked.
In my experience the problem it proposes to solve isn’t something I consider so problematic that a subscription would improve things. My experience may not be the norm, and that’s definitely a consideration I’m aware of. Still, I can’t see a reason to subscribe as such.
I saw that actually. I mentioned in another post on here recently that I figured that the only way a Fakespot v2 could exist is with a subscription model, but on the other hand, it's probably not something I could afford. Good luck with it though! You could always try advertising & affiliate links as a test to monetize the service as well.
Thanks! Advertising is certainly a possibility, though I'm not sure using affiliate links in the browser extension itself will be an option. I know Google recently changed their policy on how browser extensions can manipulate affiliate links after the Honey scandal: https://www.theverge.com/2024/12/23/24328268/honey-coupon-co...
Yeah, probably not in the extension itself. But, Fakespot, on their web based reviews, had listings to alternative or better rated products than the one you were searching for. Possibly a bit of a conflict, but as far as I'm aware, it's the only way Fakespot was monetized.
I found that it did a pretty decent job. Certainly not 100% accurate, but it often picked up on signals that made me give a closer look at a listing than I would have otherwise.
I'm sure detection is getting harder as LLMs' writing patterns become less predictable, but I frequently come across reviews on Amazon that are so blatantly written by ChatGPT. A lot of these fake reviewers aren't particularly sneaky about it.
For the unfamiliar, Fakespot was a browser extension that flagged suspicious product reviews on sites like Amazon. Mozilla bought it just two years ago and integrated it directly into Firefox as their "Review Checker" feature. Today, to my dismay, they're sunsetting it. As someone building in this space, I wrote about Fakespot's history, the problem it solved, and why we need sustainable alternatives.
Not that I'm aware of. But I do know that in late 2024, Amazon made a change where users now have to be logged in to view product reviews beyond the ones that appear on the first page (about 8 reviews). From what I can tell, Fakespot scraped the Amazon product listing pages on their backend, so that simple change would have pretty much killed its current implementation.
I can believe it. I rode in a Waymo for the first time a couple days ago and it was incredible. No problems with the rain or bad San Francisco drivers. It was a really smooth ride and I felt extremely safe.
That's just some good old Musk/Tesla propaganda. Waymo has developed a really high resolution lidar + some software magic means rain is no longer an issue for them.
They have a ton of literature at https://waymo.com/research/ and tech talks on YouTube (search talks by Drago Anguelov). They make heavy use of simulators [1] where they simulate weather events and create their own weather maps [2]. It's a very sophisticated stack.
No, ChatGPT is more than just a UI for the OpenAI API. Web requests are a feature built into ChatGPT using the API's support for function calls, but the API doesn't make any external web requests by itself.
Sorry for the naïve question, but why would I need this when most apps are distributed on Google Play? Are a lot of FOSS developers publishing on F-Droid instead of Google Play because of disagreements about Google's TOS?
Some apps don't use Google services on the f-droid variant. I guess most are publishing on both.
Neither give you an easy time. Google wants regular work from a dev (updating target SDK, filling out forms on the Play Console). F-Droid sometimes just doesn't build the app for days. (For my app it took about 8 months to initially appear on f-droid due to some bug, but hey, at least there's real people to talk to :))
