OpenAI says it has uncovered evidence that Chinese AI startup DeepSeek used its proprietary models to train a competing open-source model.
The irony isn’t lost on me as well as many others considering how OpenAI has trained its models from the web without the permission of anyone. What is that saying again? There is no honor amongst thieves.
BTW, could this (DeepSeek AI) also be propaganda? Of course… at a broader level, the same thing happens in EVERY country to get its citizens to believe in certain point of views (just as American’s have long been led to believe in this whole idea of “American exceptionalism” or having been taught sanetized versions of history).
Extreme capitalism (because of this fiduciary to shareholders line) is why much of the US’ manufacturing capabilities were moved off to China (companies even moved their customer call centers to other countries to cut expenses). American’s love of cheap crap from Walmart (often times made in China), just fuels this (American’s wouldn’t want to pay the price for the actual quality stuff that is sold domestically in China for example).
DeepSeek being propped up by the CCP would be nothing new either (to try and create doubt in the western AI market + roil the equity markets; who aren’t innocent in any of this either with their [OpenAI] exploitation of already existing content which they did not get permission to use for training their model).
Extreme capitalism led Twitter (under Jack Dorsey) to acquire Vine for $30 million in 2013 (and later shut it down in January 2017 because bottomline, they couldn’t figure out how to monetize it). While one of Vine’s original founders tried to create a separate followup (Byte), it wasn’t successful. Other companies like Instagram and YouTube did their own respective takes on it (Reels and Shorts). China’s TikTok however grabbed a load of that market (and people didn’t care about the data harvesting aspects).
CapCut (the video editing software from TikTok’s parent company ByteDance), also became heavily used (ease of use plus it makes adding effects super easy). Read the terms of service/use and privacy policy (yeah, I know most just skip past all of it) though for apps like CapCut and that should give one pause about signing up and using these apps (the other issue I have with CapCut is your movie projects aren’t easily exportable/importable to other NLE’s — a red flag because all your edits are held hostage by CapCut). The same goes for DeepSeek with its terms of service in terms of what it collects. Then there is also the security breach.
Alphabet (Google) and Meta (Facebook/Instagram) make their money off of our personal information. Instagram had its own data breach in 2019. Then there was this 2024 one (affecting Facebook, Google, and TikTok) because of how many of these companies farm out their 2FA via SMS to a third party “partner”. That’s another issue since many major corporations farm out services to other companies. Running free services isn’t really free because it’s our personal information and usage patterns that are harvested so that marketing can be specifically tailored for us and/or that data is “sold” to partners. Sure, most of their privacy policies state they do not sell it… but there is always that little provision they add about sharing certain information with partners. That is all just coded speak for making money off our personal information via some form.
The point about this/the above tangent is there is no altriusm because it’s about power and money (and trying to paint the other as the bad actor), when they are all engaging in the same practices. Would I use the DeepSeek AI app? Nope (read everything and no way was I going to use it). The other key point is the code was released as open source. That part is what is salient (as researchers are looking into that). And if there is an actual breakthrough there, so be it.
Finally, question everything. There is so much propaganda being flung around on social media (just because it’s on Bluesky with a POV that aligns with your own, doesn’t mean it is true). Similarly, my YouTube recommendations has been flooded with different takes about this (my response to that is to click the “do not recommend channel” when the algorithm begins trying to get me to click into topics like this) because I constantly “curate” to keep stuff like this plus politics out of my recommendations.
