Learn More
Resend Email
3
retoor
2d

AI is more than just a model.

It's also tooling. Tooling can help to interpret data or solve a puzzle like Sudoku or parse a JSON file perfectly. Results of those tooling will be wrapped in AI response. That quality of tooling responses is high because it's made by classical code that works with literal data and outputs literal data. As long the competition of OpenAI doesn't have tooling like that, it won't be the same.

I do assume for now that DeepSeek doesn't have that. I tried it, it answers things well, but for bigger questions that would require tooling it just crashes and says it's too busy. So I can't verify 100%.

Will try again later and update under this Rant, but assume the DeepSeek stuff is very over hyped. To know what DeepSeek really is about without watching all the fake fan video's, take this quite objective response of the maker of Perplexity AI. Someone that knows where he's talking about: (40 minutes) https://youtube.com/watch/....

So when it comes to investments of a model, what does the stuff investment is incuded in? I mean, OpenAI was way more expensive than DeepSeek but DeepSeek borrowed all OpenAI's research that was made by very expensive processes. So DeepSeek didn't pay research costs like OpenAI did. Also it (I still assume) didn't spend money on tooling.

Also, i'm sure a less woke API would be way cheaper because it doesn't have to lie to himself causing it to keep reasoning until his given woke fact makes sense! Wokism destroys models, i'm sure.

I didn't check DeepSeek on wokism yet, but it's based op GPT4o so, probably it is.

But competition is always great, I can't imagine the price would even drop further for AI requests but if it does, it would be amazing. Maybe it also becomes free and we will be forced to pay to use it without adverts.

random

competition

sudoku

tooling

deepseek r1

openai

json

api

storm in glass water

web

hype
Favorite
Ranter
Comments
  • 2
    2d
    Will soon make a blog and only mention URL to article here since random reasoning of stuff is not that highly appreciated :P
  • 2
    2d
    I do like the fact that AI is still depended on regular coding to analize literal data like statistics. But it also means that it's more a programmer than intelligent in general. Same for us that we're able to do advanced calculations using code while not really knowing the the literal math. It's kinda smart in a sense you could consider a programmer smart. Programming is an amazing way to make objective insights. In combination with its summerisation skills it does really sounds smart tho. Especially now it can doubt itself, and think a second time about a subject. Humans tend to refuse other information if their opinion is defined. For example, some give a heavy weight to information about the news paper and consider every other source as less convenient. I once told a friend a story what I've seen with my own eyes and he doubted that because there are no other sources. Wtf, does that imply that I lie or something? I worked in certain business giving me truth about X.
  • 0
    2d
    I read a good comment recently: "DeepSeek is more Open AI than OpenAI" and that is I suppose the core of the problem.

    Deep seek is just some side project of one of their engineers and It's cheaper and more available than OpenAI where they pretend It's so expensive to run these models, when in reality it was long obvious they have overblown models.

    Regarding safety, R1 is a safe model but since It's open source uncensored versions already exist.

    Also Im not sure about tooling. I remeber reading that it supports tools, but the community has issues running them right now? Something like that. But then again, It's an open source model that's relatively cheap to train. Adding tool_call examples to the training data is trivial at this point
  • 1
    2d
    ChatGPT is the part where OpenAI looses money (you are not paying the real price of AI right now). Deepseek seems to lower this cash burn which makes their business model more reliable no matter what ChatGPT does.

    It’s not the first service on that segment though, Mistral is also cheaper and faster.
  • 2
    1d
    @Tounai Hmm, while ChatGPT is actually fairly priced. I do know more people using the OpenAI api using the openwebui as frontend to spare costs tho. OpenAI is so cheap, can't imagine making profit on that. Maybe that Deepseek can do a financial analysis of OpenAI. ChatGPT never wanted to do that :D

    I do enjoy paid GPT a lot. Creating projects, uploading your project files and asking questions about it. Only thing what is rude that even when paying for it - by default EVERY TIME the 'Allow us to to train on your data' and the option is even collapsed so you don't notice it. I mean, could that few chats be so much value to them that they even do something sneaky like that?
  • 1
    1d
    @retoor
    yeah, my point is, no matter if the chatbot is better or not, if Deepseek provides even acceptable results for a low resources usage, it’s more viable than chatGPT because then it becomes affordable enough to build things with it long term.
    There’s a bunch of product on the market right now that heavily relies on GPT models which will die the day OpenAI asks them to pay for what they really use in terms of resources.
  • 1
    1d
    @retoor the moment when as a tech company you are told that now you have to stop to throw money out of the windows is really a tough moment that isn’t the funniest thing ever.
Related Rants
devRant © 2021 Hexical Labs LLC
Privacy Policy  |  Terms of Service
Add Comment