Відео

Guys, you've been in Shanghai for many years almost decade. Can u tell us more about Chinese mindset and strategy how they became Tech Giant and now competing US. And also let us know pros and cons of it in terms of Western economies.

Thx for the fair review and introducing Cartario to me! Used Dall e from time to time via bing. Will check the new options soon. Impressing how much is changing within a couple of month ❤❤❤😅

You asked for a she lion, not a male lion????

They both failed the lion image. "She is making an angry expression", but they are images of a male lion.

Didn't your lion prompt say "she"? If so, both image generator produced male lions. Not very smart of them, methinks.

The version is still staging, no official announcement on DS homepage.

Wait, Deepseek spoke English and ChatGTP spoke Russian! LOL.. that's hilarious! I wonder why?

Good enough for me already.

Check out www.cartario.com to create the best AI images. You can also train a custom AI image model on photos of yourself, your products, your pets, or whatever you want! Try it out.

It's free ! Test again after the release "age" same as OpenAI

Its just beginning. So i won't see it as failure

Agree. I tied it to, the images look like ass.

If the U.S. wants to win the AI race, it needs to stop giving data to China. TikTok collects vast amounts of data from Americans daily, which can be used to train its machine learning algorithms. China relies on U.S. data because its domestic data is heavily one-sided, primarily focused on praising the CCP. In contrast, American data is more diverse and balanced.

nice test, willing to see more

hahaha. You were wrong. You never underestimated the Chinese AI.

SO FAR, the English were good. perfect grammar and lots more

Try asking Deep Seek about anything negative that has to do with china or even factual things like the ccp. If they sensor all of that, how can we trust how they trained the data about other historical events like it may be prone to describe the west as the enemy because they're so skewed in what data its fed. Logics and math may be great but I would not use this for fact checking plus anything they say its build on without us having full insight is total bullshit, they've never been truthful about anything you can check yourself.

The problem for Nvidia is that it's products will become commodity soon. Chinese will be capable of beating not only best LLMs, but also beat chips made by "America". So will will have glut of AI training chips soon from multiple sources and Nvidia has no moat.

Super Intelligence is about to come!

they are large LANGUAGE models. If you want a mathematical model you might be better to use RBES than NN.

As a gamer, I don't care the AI, I hope I can get a cheaper Nvidia GPU for gaming!

If anyone loads this on their device, you are insane. This is the mother of all viruses/hacks.

Awesome video, please check out objectivism, art is not subjective, but objective, you can judge art by standards, you can give it points based off objective aesthetics. Art can represent metaphysical, epistemological and ethical facts perceptually to a human. LiquidZulu on UA-cam has a great video on this (in defense of AI art) and Ayn Rand has written the Romantic Manifesto on it which I have yet to read. Keep it up!

Check out our custom AI image generator. You can make a custom image model of yourself, your products, or even your pet! www.cartario.com

Unfortunately it appears, that the conversion of the sudoku picture into the text was also erroneous. And the whole exercise here is worthless.

By now everyone must have realized that the deepseek sudoku numbers are complete bogus in that the given numbers have mostly changed. Not only are duplicates in a 9 sqare box but the whole game was turned upside down.

Rumors are they have tens of thousands of h100 but just not saying it cuz they had to smuggle them

Some one asked Deepseek what model are you? Deepseek answers it is ChatGPT. 🤭

My dude, Deepseek did not copy ChatGPT. It is a new model and very innovative

I simply don't trust China and their products, I can bet soon some shady thing about it will pop out.

Does this means china going to win ai race?

I see a deepseek back coming soon.

I think chat GPT 4 is more effective than deep seek. The only benefit of deep seek is free access.

I heard that what Trump spent for the mexican wall, China spent similar amount to land its rover on the far side of the moon. 😮

I tried a few Sudoku puzzles. It seems Claude 3.5 is no doubt the best. It answers all the questions without any issues in less than 2 min a questions. ChatGPT o1 can answer the 1st one but not the advanced puzzles. DeepSeek R1 cannot even solve the easiest one.

What are your thoughts on using Deepseek for building codes that will be used for creating resources(like ec2 instances) in AWS? Kinda worried asking it because I was told that Deepseek was made in China.

In my opinion they have worst chip but give users the ability to see thinking of models to get better performance

Nowadays China has better engineers while Americans get dumber

Stop crying about machines not doing what you want. You’re delusional if you think your whining matters. The machine doesn’t care about your fantasies. Deal with it or shut up-either way, you’re still wrong. Look at these crybabies whining when life forces them to get better. They’d rather stay useless than admit they need to be dragged out of their pitiful comfort zone. Stay mad-it’s all you’re good for.

Nice video. Everything is clear.

business model of OpenAI is fucked. HAHAHAHA why use them when i can put a Free, unrestricted, uncensored OpenSource version in my basement

The reason is chinese and russian are good in technology related things American can be hypocrite or racist but that the truth in game like pubg and more even game engine like unreal and Many popular title and technology you use because there are people with a brain the size of a peanut 🥜 that don't understand the reason why China is always winning the race even if they join the race late think 🤔 your self and ask the question don't be hypocrite for real this time

Does that mean you are able to run this model offline and that you own it and no one can take it away from you? What would the benefits of that be versus using one online.

Seem like the text version paste into DeepSeek and ChapGPT doesnt match the soduku image. If he used a chatbot to convert the image to text, it was incorrect.

Open source is awesome

Turn off this video after hearing “Knows little bit English”, “does not have the resource to tweak it”

I enjoy your grounded delivery, a focus on the facts and your insights. Subscribed.

America needs to stop sanctioning China. Soon windows will be replaced and I can bet on that. They already have alternative for MS Office.

You can train the model better by replying to Deepseek, the answer is wrong, one of the 9 numbers matrix boxes is wrong. Thru this feedback, you help to get it to improve, thus benefiting the rest of the users

I don't understand why you say OpenAI can't be more efficient, the methods are in the scientific papers they released and the models are open source. They have enough smart people left at OpenAI that can do that I believe, or at least some of the western competitors can. The real innovation that very few are talking about is how they did it and I'm no expert, but it's how they combined Reinforcement learning (RL) for creating LLMs. 3:11 this is probably in part BECAUSE of the sanctions, Chinese knew they had to do with less so they focused on making it more efficient. 3:50 yes and I do think this was always known to happen, the sanctions are only a delay tactic, never a full solution. But I think they underestimated the Chinese ability to innovate and how much can be improved if you focus on efficiency instead of everything else. I also do think the Chinese were happily surprised by the results they got, probably worked even better than expected (kind of get that from the tone of the paper of the few parts I've seen of it).