Posted on 01/30/2025 7:25:05 AM PST by zeestephen
CNBC asked industry experts for their views on DeepSeek, and how it actually compares to OpenAI, creator of viral chatbot ChatGPT, which sparked the AI revolution.
(Excerpt) Read more at cnbc.com ...
DeepSeek is built on other peoples work , typical Chinese ,LOL
Extraordinary claims require extraordinary proofs..........
DeepSeek’s AI claims have shaken the world - but not everyone’s convinced
—
Especially Alibaba who claims their AI is far superior to DeepSeek.
One thing is for certain, you will not get an unbiased article from the US press. They will shill for ChatGPT as if their lives depended on it. They know who controls the narrative.
Try this for starters:
Fourteen Impressive Examples that Show Deepseek is Better than ChatGPT
https://greekreporter.com/2025/01/30/deepseek-better-chatgpt/
AI exec on Fox last night said they trained DeepSeek using ChatGPT.
When DeepSeek was asked “Who are you?” It replied “I am ChatGPT.”
When DeepSeek was asked “Who are you?” It replied “I am ChatGPT.”
“I am Spartacus!”
DeepSeek is built on other peoples work, typical Chinese
At any event, it’s about 95% as ‘useful’ as ChatGPT, and can be dropped into a computer that doesn’t have access to the Internet. This means the Chinese will only be able to spy on entities that run DeepSeek using internet-connected servers.
I can build an AI very easily.
All I need to do is you submit questions to YoYoChat, then my back office program will submit the same question to ChatGPT. I then relay the ChatGPT response to you as my YoYoChat response.
Someone said that when they asked DeepSeek ‘who are you?’, it replied “ChatGPT.”
Hey, did they just steal everyone else’s software then only spend $5.6 million to reassemble it into DeepSeek? That’s more like their modus operandi.
Seems like the perfect False Value scam
Tom’s Guide did a nice comparison with real world examples, though you won’t find CNN or MSNBC covering this. It should be pointed out that if DeepSeek totally copied ChatGPT the answers would come out the same or at least similar.
I tested ChatGPT vs DeepSeek with 7 prompts — here’s the surprising winner
https://www.tomsguide.com/ai/i-tested-chatgpt-vs-deepseek-with-7-prompts-heres-the-surprising-winner
The biggest problem with ChatGBT is they used lowest common denominator and biased sources to train it. No one should be surprised that it spits out answers that are lowest common denominator and biased. The #1 rule of computing still holds true: Garbage in, garbage out.
LLM’s will soon be a dime a dozen if they are not already. They are a commodity now. There is no money in developing a LLM. The money is in the application layer of this industry. Putting LLM’s to work to build AI models that are attached to a business function to improve something. That’s the secret sauce. Not the LLM
When your LLM is trained on reddit, there is a problem.
It would be the same if you trained AI in a public school classroom. AI would be as illiterate as the bulk of American schoolchildren. Imagine an LLM trained solely on the writings of the greatest thinkers of the ages, Greek, Roman, and Asian philosophers. That would be worth using. Instead we get Artificial “Intelligence” based on three letter agency bots and biased political shills. It will never be ready for Prime Time.
If the cost savings of DeepSeek are real, it is possible to create a model that is based on the deepest thinkers of humanity through the ages and apply that to advance human goals. Anything else would then be regulated to spitting out writing assignments for school children that keep leftist teachers’ happy their propaganda is working and tracking of your grocery list on your smart refrigerator.
It is stolen tech. The Chinese think nothing of theft. They couldn’t build an “Eliza” on their own today much less today’s AI.
Every key stroke you make on that function can be read by the Chinese communist party.
I agree with you I was not trying to comment on the quality of the LLM’s, just their place in the A.I. ecosystem. As always garbage in garbage out. there will soon be hundreds of them, and a few will be good, the rest will just fade away.
The problem is they won’t fade away fast enough and will be gumming up the works while they are here. Most business models appear to be nothing more money laundering operations.
And then there’s the whole security thing... https://www.wiz.io/blog/wiz-research-uncovers-exposed-deepseek-database-leak
Claude is my LLM of choice.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.