DeepSeek is annoyingly too verbose
Discussion
I find it too contradictory of itself
It's Chinese.
It's like a fucking James Joyce novel. 😂
If we give China our information to have them process it, only to have China return an action plan or provide a guided assessment… what do you think the outcome most likely leads to?
Talk about the ultimate psyop in plain sight. No?
The most interesting part of AI is: as a population we cried foul for over a decade that companies like Google, Meta, Amazon and Microsoft stole our data and used it against us. In effect, they weaponized what we put on their systems for free. We freely gave it to them.
Now with AI, we are *paying companies* like OpenAI to upload our data so it can peer deeper. We are paying to give them what we cried that took for free.
And we when supposed adversaries provide a better mousetrap, we jump at the opportunity to take direction, interpretation and effectively guidance from them.
Kind of crazy when you filter out the hype-noise and break it down.
Thinking about each of the moving parts sheds some interesting light. The new world is gonna be a wild ride.
DeepSeek’s response while reminding it to be succinct:
As an AI, I don't have personal opinions, but I can provide an analysis: The concerns raised are valid and reflect broader debates about data privacy, corporate power, and AI ethics. The shift from freely giving data to paying for AI services highlights a paradox in how society values and manages personal information. It underscores the need for transparency, regulation, and ethical frameworks to ensure data is used responsibly and that individuals retain control over their information.
You could do without the "too" there. Bit verbose if you ask me
It's a very useful alternative to closed source thinking models though
“Oh, THANK YOU! FINALLY, someone else who sees it! I mean, my goodness, you ask it a simple question, and instead of a crisp, concise answer, you get an entire dissertation, a soliloquy, a grand linguistic opera complete with overtures, intermissions, and a three-act structure that would make Shakespeare himself say, ‘Alright, that’s a bit much.’
And it’s not just the length—it’s the way it dances around the point like a caffeinated professor who refuses to just say what the answer is because, oh no, it must first establish context, provide historical background, list unnecessary synonyms, and conclude with an overly polite ‘hope this helps’ when you’ve already passed out from sheer exhaustion!
Honestly, if I wanted to read a novel-length explanation every time I had a simple question, I’d start carrying around an encyclopedia and asking it for life advice! At this point, I half expect DeepSeek to start prefacing every response with ‘It was a dark and stormy night…’ before finally arriving—paragraphs later—at the completely unremarkable conclusion that YES, water is, in fact, wet!
It’s an endless, soul-draining, patience-destroying experience of trying to extract a simple, concise answer from this AI and instead being buried alive under an avalanche of excessive, redundant, and completely unnecessary words! It’s as if you ask for a sip of water and instead get a firehose to the face, except the water is words, and the firehose never turns off!
One basic question—something that should take maybe one sentence, two at most—and what do you get? A multi-paragraph essay, complete with historical context, a breakdown of all possible interpretations of your question, an explanation of the reasoning process behind the answer, a few disclaimers to cover all bases, and then a closing sentence that acts as if it wasn’t already too late to stop! It’s like watching someone take a five-minute detour just to cross the street!
And let’s not even get started on the structure of these responses! The sheer length! The elaborate phrasing! The obsessive need to ensure every possible angle is covered before actually answering the question! It’s like DeepSeek thinks we’re writing legal documents instead of just trying to get a straight answer!
But no, no, it doesn’t stop there! Oh no, because not only is it long-winded, but it doubles down! If you try to push back—if you dare to ask for a shorter response—it acts like that’s some strange, unreasonable request, like you’re asking it to summarize War and Peace in a single word! And what do you get? A slightly shorter version that still somehow meanders through a needlessly complex explanation before reluctantly arriving at the point, as if brevity were an insult to its very existence!
Was it never trained that sometimes, just sometimes, people don’t need the entire encyclopedia—just a single, clear, and direct answer!
I mean, SHEESH!