Incidentally, a very fun read on the dynamics behind the whole thing https://americanaffairsjournal.org/2021/08/the-value-of-nothing-capital-versus-growth/
What ultimately matters is the algorithm that makes DeepSeek efficient. Models come and go very quickly, and that part isn't all that valuable. If people are serious about wanting to have a fully open model then they can build it. You can use stuff like Petals to distribute the work of training too.
Turns out that when you kidnap people off the street and gang press them into fighting, they're not terribly motivated.
The code is open, weights are published, and so is the paper describing the algorithm. At the end of the day anybody can train their own model from scratch using open data if they don't want to use the official one.
What you said is demonstrably nonsense and I very much did link sources to the model, and the paper explaining how it works, and why it's open source. Absolutely hilarious that you keep lying here while accusing me of acting in bad faith. You are an utter clown.
Oh no, people here have biases different from your own. The horror! Meanwhile, I provided sources in other responses in this thread. You could also just spend your time to google this stuff yourself instead of trolling here, it's not exactly hidden.
Bye!
Pretty much all my interactions with the community here have been positive, aside from a few toxic trolls such as yourself. Maybe take your own advice there champ.
I have backed up my claims with evidence and sources. Stay mad troll.
Exactly, in the west competition between companies is a zero sum game. This precludes companies from cooperating on technology, and leads to a lot of duplicated effort as people reinvent the wheel over and over. In China, competition is often a positive sum game where everybody gets to learn from each other, and move forward together.
Seems like most people desert or surrender at the first chance they get. Sometimes they just don't get an opportunity to do that before they get to combat.