News DeepSeek.com making waves!

k660 · 2025-01-31T21:54:04+0530

meme -
(voice - oversimplified )
hooman: inbreeding is bad
ai: i am about the end entire career of those who believes that (i am also indirectly implying bro,ai, crossed line just to end human life/humans as we see)
hooman: wai... what?
ai: if done correctly it can result in good too ( then proceeds to show how ai changed from 2021 to till now present )
hooman: i think that is enough internet for today
ai: you think i am limited to only internet... huh.
hooman: we are ****ed arent we?
ai: yes (immediate answer)
hooman: i didnt ask you!!
ai: proceeds to delete his answer
hooman: wait when did you get that update?!? (O_O)

(you know while typing this meme, it turned into a small video script somehow)

TEUser2K1 · 2025-01-31T21:54:16+0530

Emperor said:
As after so many years, every one need to upgrade, syllabus too

That part they are doing by themselves, probably don't need direction advice from those already failed miserably.

Kaching999 · 2025-01-31T21:55:15+0530

Emperor said:
@Kaching999
hey don't loose patience, every one trying to explain at their best, some one learn quickly some later, thus errors & omission should accepted, please do not stop posting/replying, if you think what you post is correct keep firm on your ground, after all that is essence of this forum.... even non tech person like me learns lot of things more than I doubt I could if join any PAID crash course

chill bro all is good I will still post in thread but I won't be banging my head against a brick wall unfortunately. All I want people to understand that a publicly available open source software is always better especially when it is an AI model which you have the option to run locally on your system. When thing get too political for me I don't want to continue in that direction there are nuances to such situations which need a person to look at both the sides but when someone is already heavily biased towards one side. I have already said my piece in the previous posts and don't feel like repeating myself again and again. I have evaluated both the sides and stand firm on my belief and understanding, like Louis says "Never accept the premise of...", well you might know that one already.

blr_p said:
No, I have argued my points wuth examples all throughout. I see nothing wrong with what I have posted at all and neither are you capable of showing it.

You are just ignoring what I said and repeating your same points. And blaming me for challenging you.

sure buddy, whatever makes you sleep at night.

blr_p · 2025-01-31T22:06:28+0530

Futureized said:
No AI model can work on its own, it has to feed itself to gather as much content as possible, whether it be offline/online obtained legally/illegally copyrighted or not.
Its atleast similar like redefine Search Engines with advanced features.

A very interesting point.

madhukannan · 2025-01-31T22:12:31+0530

nullpc said:
It's a reasoning model, this is expected. What you've posted is what the model is "thinking". The actual answer is below </think>

There was nothing. Now I asked the same question its telling me to refer govt websites or news websites

blr_p · 2025-01-31T22:13:51+0530

Kaching999 said:
Unfortunately, Open AI is not a community project but a government and private funded multi billion dollar organization that only works for profits, if you can't separate that from an open source piece of software there is something wrong there.

Open source piece of software controlled by China IS what is wrong here.

Do you trust them? Just because this is open source

Skoomahed · 2025-01-31T22:17:16+0530

blr_p said:
Open source piece of software controlled by China IS what is wrong here.

Do you trust them? Just because this is open source

Open source makes it easier to see the inner workings of the model. The added transparency is why open source in general is popular and widely supported. If you had to choose between a proprietary model like gpt and open source like the R1, it's pretty clear which is the easier choice to make.
Being able to run r1 locally means your data is safe with you.

TEUser2K1 · 2025-01-31T22:18:41+0530

blr_p said:
Do you trust them? Just because this is open source

No need to trust that s/w, one can execute the s/w in a sand-boxed, isolated environment and learn how it behaves.
If it's opensource, it's one's duty to at least audit the code and utilize it. Have to expect this minimum when millions and credit worthiness of the country is involved. These kinds of questions shouldn't even arise...

blr_p · 2025-01-31T22:19:37+0530

Skoomahed said:
Open source makes it easier to see the inner workings of the model. The added transparency is why open source in general is popular and widely supported. If you had to choose between a proprietary model like gpt and open source like the R1, it's pretty clear which is the easier choice to make.
Being able to run r1 locally means your data is safe with you.

I don't argue about the benefits of open source. Well known for over twenty years.

What do you do if they go closed source?

Fork right. Except this time lots of funds needed for a working alternative.

TEUser2K1 · 2025-01-31T22:23:27+0530

blr_p said:
What do you do if they go closed source?

Fork right. Except this time lots of funds needed for a working alternative.

The entire activity is to be self dependent after learning from the code, not perceptually dependent on others.

Kaching999 · 2025-01-31T22:26:19+0530

blr_p said:
I don't argue about the benefits of open source. Well known for over twenty years.

What do you do if they go closed source?

Fork right. Except this time lots of funds needed for a working alternative.

Now who is talking about hypotheticals

TEUser2K1 said:
The entire activity is to be self dependent after learning from the code, not perceptually dependent on others.

Literally

blr_p · 2025-01-31T22:28:55+0530

Kaching999 said:
Now who is talking about hypotheticals

Can you think of a worse case scenario?

Considering this is an adversarial power we're talking about I would argue its credible.

Kaching999 · 2025-01-31T22:31:49+0530

blr_p said:
adversarial power

buddy nobody can hack you if you run the model locally, do you even know how an open source AI model works?

blr_p · 2025-01-31T22:31:56+0530

TEUser2K1 said:
The entire activity is to be self dependent after learning from the code, not perceptually dependent on others.

And when you no longer have access to the code what happens?

Your ability to learn remains at the same level.

Meanwhile China has got the world to work for it and take things forwards from there.

Development here requires world collaboration. What China does with the fruits of this labour is the threat

Kaching999 said:
buddy nobody can hack you if you run the model locally, do you even know how an open source AI model works?

Obviously I do. Hack is not the problem.

Assisting an adversarial power is the problem

Kaching999 · 2025-01-31T22:33:10+0530

blr_p said:
And when you no longer have access to the code what happens?

how is that even possible again, do you know how open source software works and progresses??????

mayank11280 · 2025-01-31T22:34:07+0530

Umar0039 said:
It's good but not that great, same with ChatGPT. Once you start asking deep technical questions it hallucinates.

You can start asking me deep technical questions, and I will hallucinate.

Kaching999 · 2025-01-31T22:36:53+0530

blr_p said:
Assisting an adversarial power

and you think that USA which has famously destroyed the multiple governments only because they opposed them is not an adversarial power??

blr_p · 2025-01-31T22:37:27+0530

Kaching999 said:
how is that even possible again, do you know how open source software works and progresses??????

In the cnbc video arvind mentioned it could go close source.

So you remain at whatever last build you have. I'm not saying you lose that code. Just that you're on your own thereafter ir whatever collaboration coalesces around the fork.

Kaching999 said:
and you think that USA which has famously destroyed the multiple governments only because they opposed them is not an adversarial power??

The US is not threatening India. Neither are we threatening the US.

But China is a threat to both the US & India.

How do you handle a bully like China. Make friends with an even bigger one. Thankfully they're not communist this time

Kaching999 · 2025-01-31T22:42:38+0530

blr_p said:
In the cnbc video arvind mentioned it could go close source.

So you remain at whatever last build you have. I'm not saying you lose that code. Just that you're on your own thereafter ir whatever collaboration coalesces around the fork.

An communities all around the world have done much better than large closed source software corporation on their own, Libre office, adblockers, linux, Apache, unity, source need I go on?

TEUser2K1 · 2025-01-31T22:45:12+0530

blr_p said:
Your ability to learn remains at the same level.

If that is the case, we declare ourselves dumb and perceptually stop doing things /s. And guess what you got company !

Infosys founder Narayana Murthy on why India should not build its own AI models: “...the Indian mindset is still not…” - Times of India

TECH NEWS : Infosys founder NR Narayana Murthy has sparked a debate by questioning India's pursuit of homegrown large language models (LLMs). He argues that India

timesofindia.indiatimes.com

Dude, we are learning to develop our own AI, by combining ideas from other AI tools, for some of which even the source code is open. What more do you need !

blr_p said:
Considering this is an adversarial power we're talking about I would argue its credible.

If India is an independent country with it's own identity, everybody else are either adversarial or at least indifferent. Rest are fare weather friends. Talking as if USA suddenly is best phriend, wen dtrump just haven't settled his tongue after threatening brics and those who trade against usd.
No need of AI in self writing jokes here