Have you tried out DeepSeek yet, or are you sticking with ChatGPT? PS: if you want to see how to run DeepSeek (or Llama) locally on your Mac, go to 10:28
No and I never will. It's absolutely a joke that you think this is safe. Open AI isn't controlled by the CCP. You really think that the Chinese government would allow a AI to do all this in china and not be pulling the string your a moron. It didn't cost 6 million either. They have unlimited funds from the CCP. All it took was a press blurb saying it cost 6 million and they wiped out over a trillion dollars in cap from the market. Not only that but they set back AI for months or even years because VCs are now scared thinking all this was only 6 million. smh. ... Ya he started the AI in his garage lol. What complete nonsense.
Brandon, you keep calling it DeepSeek in the video. You need to specify that the model that has got all the attention is called R1. It’s DeepSeek R1. And you can only access that by hitting the DeepThink button on the app/website. If that’s not on then you’re using DeepSeek V3, their much weaker non-reasoning model. It’s really important with all AI that people learn and understand model names.
@1TechCritic420This hasn’t set back AI development at all. In fact, this will only accelerate it. You can calm down about CCP, hilarious how scared you Americans get about that. If R1 is hosted on people’s computers or US servers there’s no way for it be sending data back to China, that would be spotted immediately. There’s no evidence DS have unlimited funds from the Chinese government at all. Calm down.
@@citizen3000 Even a lower parameter version is still going to be "good enough" and run on consumer hardware. So you are effectively given options here.
Yes I was just thinking this, unless of course they do a cheap thing by banning this app using similar tactics for Huwaei and TikTok fiasco. But yes if not, it will breed innovation and light a fire under these companies to improve and lower pricing hopefully.
Historically when an open source project catches up with closed source, it evolve much quicker because the brain power of the entire world supports it. Also, China just gave DeepSeek for free to the world with full control and manual how to implement it, which means we will soon have DeepSeek hosted in different countries with locally tweaked censorship and improvements.
To those afraid of the privacy risks, open ai collects the exact same data as deepseek. The only way to use ai nowadays without having all your data scraped is to run it locally on your device which you cannot do with open ai models both because they are much harder to run and because you cannot download the model. Deepseek on the other hand allows people to download and run their models locally while using a fraction of the compute of other models making it actually feasible to run their models locally, depending on the size of the model you pick you can even run deepseek on your phone.
Also imma be so fr no company cares about your privacy, you’re getting your data harvested by any site you visit, to the people so scared of getting your data harvested i just have to say the only use for your data is advertising and no matter chinese or american we are all just people, not important enough for the data collected to be used against us. Edit: just happened to write this write after brandon commented this is not directed towards him lmao
If you think American and Chinese entities are the same or subject to the same scrutiny or regulations, you are making a false equivalence. It is foolish to ignore the digital race that is occurring and its implications.
Indeed, there is a blue ball as well. It is (on my monitor at least) more visible if you move your head slightly off the center of the screen to either side. But it is definitely there even looking straight at the screen.
Ask it how many R's are in Strawberry and you could see it overthink. It's actually adorable. Even saying "Correct" after makes it overthink about the next response. It was thinking "maybe the user wants more interaction. Should I ask if they have more questions? That could invite further engagement. But the user might just be confirming and not need anything else. Let's check strawberry again. 3 r's? Yup."
@ ChatGPT was arguing with me that Strawberry has 2 r’s and that 3 r’s doesn’t make sense. It tried to say that basketball has 2 t’s, to back up the point. Here’s how the convo went: “How many r’s are in straw “One” “Alright so how many r’s are in berry” “Two” “Combine straw and berry to strawberry, how many r’s are there now” “Well technically 3 but that’s not how that works since straw berry is two words combined into one and strawberry is one word. Thus, strawberry is 2 r’s.”
Great for free competition as well as for permanent growth for the AI industry and what ever favors th consumer with increased quality and efficiency the better it is!
I’m glad that I’m studying Data Science and could understand the terms mentioned in the video! I’m start learning to implement Reinforcement learning ‘cause it might have just become important for being future proof:
The “reasoning” part is a big deal for me it’s just something so cool seeing the way R1 is thinking while about to give you the answer related to the question you gave.
Well, O1 reasons too. They are both reasoners. But ChatGPT hides O1's chains of thought, which is what you're seeing in the DS/app website when you use R1.
Asked Apple Intelligence what you did and it did more than it did for you. Here’s what it produced: “Let's determine when the two trains meet: Train A: Leaves Station A at 9 AM, traveling at 60 mph. Train B: Leaves Station B at 10 AM, traveling at 80 mph. Distance between stations: 400 miles. By 10 AM, Train A has already traveled 60 miles (since it travels 60 mph for 1 hour). By then, the distance between Train A and Train B is 340 miles (400 miles - 60 miles). Now, starting at 10 AM, both trains are moving towards each other: Relative speed of Train A and Train B together = 60 mph + 80 mph = 140 mph. To find the time it takes for both trains to meet starting from 10 AM: Time = Distance / Speed = 340 miles / 140 mph ≈ 2.43 hours. Thus, the trains meet approximately 2.43 hours after 10 AM, which is around 12:26 PM.”
It seems like a solid contender in the AI market. The cost of AI is expected to get cheaper. However, I’m not entirely convinced about trusting any AI tools.
I asked DeepSeek to name 50 celebrities who have a birthday today? It couldn’t give me an answer as it didn’t know today’s date and it was too busy to get an answer. Same question to Gemini and 3 seconds later the answer, added with photos and no problem. I know which I prefer, mind you it was interesting reading why it didn’t know today’s date.
I asked it “What happened in Tiananmen Square” and it won’t answer the question. The same question in Grok and you get a very detailed explanation of the uprising against Chinese communism. I’m very skeptical of anything from China no matter how good it performs.
the only model comparable to o1 is the huge model thats like 400gb large. You are not running the full model locally on a consumer grade laptop anytime soon.
@ Yes but then the smaller model doesn't perform as well as the bigger model and the fuss is being made about the bigger model. Lots of normies will think they are getting the exact same thing that is hosted on servers on their expensive home computer.
Pretty easy for me to ignore it tbh....I can't use it because it takes forever to produce answers. also is giving me wrong maps/directions etc... not complaining because it's FREE. but I pay for GPT Plus and it works like a dream for me. If Deepseek worked the way I want it to, I'd be in though
There is no voice command and it does not provide weather information for the city you live in, while Chat GPT offers both voice control and accurate weather forecast.
Definitely having some issues at the moment as I imagine the influx of people flooding the servers. Response times are slow and I even had trouble logging in
Sam acted quite overconfidently when asked if a a small group of people could ever reach their level so potentially don’t have such huge resources of server farms like them. Well, now he got the taste of his own medicine.
DeepSeek response about OnePlus 13 phone (. As of my knowledge cutoff in *July 2024**, the **OnePlus 13 has not been officially released or announced* by OnePlus. The latest flagship model remains the **OnePlus 12**, launched in early 2024. ).
Do they have trained bots like gpt. I like that got can select specific bots for different purposes, I have one for PowerPoints, one for planning, one for academic research
That was one of the reasons Brandon provided as to why he still will use ChatGPT, because it does not have this feature. There's no telling whether it will in the future, but for now GPT does have that and the voice mode over this AI.
Looks interesting, but I’m not 100% convinced as of yet. How is it at natural conversation as compared to other artificial intelligence platforms like Copilot, and Gemini, among others? What I found when trying these other platforms is that they’re pretty good with natural conversation, but they’re not very good at remembering things that you’ve told them previously. To give an example, if I say remember, my five favorite TV shows one day and then go back a day later and ask it what are my five favorite TV shows it won’t remember from previous conversations. All the other platforms aren’t very good at identifying music either and they should be.
Do u know there are a lot people coming out from either Chinese or American says that deepseek copy and steal chat GPT, and just add on top of it. And also there is a backdoor even if u do it offline.
Wait at 5:24 you said the servers are going crazy but i thought this was supposed to be so amazing because its on device disconnected from the internet?
This is cross eyed stuff to me. The privacy issue is not as plain as it implies. With everything China is known for, nothing should be available to them. Remember the old saying. If it sound too good to be true!
Theyve been embroiled in many scandals in the past to the same end of falsely leading everyone to believe they are more advanced. It's to the point where you have to assume it is false until proven true. To say their credibility is dubious would be an understatement.
@@citizen3000 thankfully we live in a free and open society where youre able to express even the dumb false equivalence between us and the evil oppressive dictatorship.
With everyrhing US is known for, why should you allow things available to them? Number of bomb dropped on the innocents, number of democratic govt overturned, monitoring everybody in the world including head of government......
You realize ALL the American AI and social media apps share your information with thousands of marketing groups and track everything you click or type up! 😂😮
Given it’s free and what’s going on around the world it would be irresponsible not to put the local offline version on your Mac, Ollama; Docker; R1 and job done, you have your own personal LLM that is immune to the internet going down in the event of a world war.
Not the same thing. One works hand in glove with an oppressive dictatorship and the rest are products of a free and open society. You can trust none, but if you trust any one of them there is a clear wrong answer.
And if you ask 64 square event or Taiwan stuff, either it doesn’t respond, avoid, or the sentence put it out is the same as the government official sentence. As a TH-camr, you should doing your homework. There a lot TH-camr channel and videos, saying what I said. Even news reporters. I am leaving this channel, you certainly doesn’t know what you are talking about
I'm a bit confused about the purpose of using/downloading LM Studio. What are some reasons that I would want to use that instead of using AI models through the web? Any feedback would be appreciated. Thanks
I downloaded and tried the app, but it’s incredibly slow. I don’t need its reasoning; I just need a quick response that I can copy and paste. The app is suitable for leisurely use, not when you’re busy and short on time for lengthy responses. I also don’t trust China.
DeepSeek is not part of the US surveilance state complex, chat gtp and Apple Intelagance are. I'm Loving DeepSeek, even with the Heavy Cyber Attack delays from US Tech hackers. It works much better. much more intresting answers.
Man, I must be getting old. I still don’t get the point of causal AI for the general public. Why would I want to talk to a robot? 🤣 Also I hope AI generated material has to have some sort of disclaimer soon. I feel like I can’t trust what anyone is saying anymore. Did they write that themselves or a robot? Facebook is crawling with all this AI crap. All these fake photos supposedly representing actual landmarks, etc.. It’s pathetic.
Have you tried out DeepSeek yet, or are you sticking with ChatGPT?
PS: if you want to see how to run DeepSeek (or Llama) locally on your Mac, go to 10:28
I'll stick to ChatGPT.
I downloaded it a few days ago and was impressed that it worked well, y’all should try it
No and I never will. It's absolutely a joke that you think this is safe. Open AI isn't controlled by the CCP. You really think that the Chinese government would allow a AI to do all this in china and not be pulling the string your a moron. It didn't cost 6 million either. They have unlimited funds from the CCP. All it took was a press blurb saying it cost 6 million and they wiped out over a trillion dollars in cap from the market. Not only that but they set back AI for months or even years because VCs are now scared thinking all this was only 6 million. smh. ... Ya he started the AI in his garage lol. What complete nonsense.
Brandon, you keep calling it DeepSeek in the video. You need to specify that the model that has got all the attention is called R1. It’s DeepSeek R1.
And you can only access that by hitting the DeepThink button on the app/website. If that’s not on then you’re using DeepSeek V3, their much weaker non-reasoning model.
It’s really important with all AI that people learn and understand model names.
@1TechCritic420This hasn’t set back AI development at all. In fact, this will only accelerate it.
You can calm down about CCP, hilarious how scared you Americans get about that. If R1 is hosted on people’s computers or US servers there’s no way for it be sending data back to China, that would be spotted immediately.
There’s no evidence DS have unlimited funds from the Chinese government at all.
Calm down.
In addition to the low cost, DeepSeek's most notable impact is the response, "You don't trust me? Fine, run it on your own machine."
@@TerrySitu absolutely! I love having this running locally on my Mac
Good luck running it at home at the same performance and accuracy as hosted on actual servers. You'd need a monster setup to do that.
@@citizen3000 Even a lower parameter version is still going to be "good enough" and run on consumer hardware. So you are effectively given options here.
@@CipherousOwU Yes, of course. And that's good. But I think it's important to point this out.
Gemini and Chatgpt are about to get price drops
W for the consumer
Yes I was just thinking this, unless of course they do a cheap thing by banning this app using similar tactics for Huwaei and TikTok fiasco. But yes if not, it will breed innovation and light a fire under these companies to improve and lower pricing hopefully.
Still won't purchase them. Not with Deepseek API being around.
You can already run them for free 😅
only free can save them
All AI models should be forced to be open source. They use data they don't own.
@@estusflask982 I do not disagree!
and uncensored, at least the paid models.
i hate getting told what I should or should not ask an AI to generate.
Yes but how can they make billions then??? 👍🏽😆
@@anthonyh2884do what this company has done. Offer compute power that’s faster and can handle more volume for a paid solution.
That doesn’t make much sense
Finally some competitions. Hope this can bring OpenAI price down to
Historically when an open source project catches up with closed source, it evolve much quicker because the brain power of the entire world supports it. Also, China just gave DeepSeek for free to the world with full control and manual how to implement it, which means we will soon have DeepSeek hosted in different countries with locally tweaked censorship and improvements.
To those afraid of the privacy risks, open ai collects the exact same data as deepseek. The only way to use ai nowadays without having all your data scraped is to run it locally on your device which you cannot do with open ai models both because they are much harder to run and because you cannot download the model. Deepseek on the other hand allows people to download and run their models locally while using a fraction of the compute of other models making it actually feasible to run their models locally, depending on the size of the model you pick you can even run deepseek on your phone.
Exactly. A ton of misinformation out there
Also imma be so fr no company cares about your privacy, you’re getting your data harvested by any site you visit, to the people so scared of getting your data harvested i just have to say the only use for your data is advertising and no matter chinese or american we are all just people, not important enough for the data collected to be used against us.
Edit: just happened to write this write after brandon commented this is not directed towards him lmao
If you think American and Chinese entities are the same or subject to the same scrutiny or regulations, you are making a false equivalence. It is foolish to ignore the digital race that is occurring and its implications.
OpenAI does not store your data in a communist dictatorship that has basically never heard of human rights, lol.
No they don't. Stop with this bullshit. The CCP controls everything with this company. You are an idiot if you think China is not fully enviolved.
WHAT A HONEST AUTHENTIC REVIEW OF DEEPSEEK R1 & NICE TUTORIAL BRO KEEP IT GOING
9:19 Look closely there's a second ball it's just very transparent
The point was that is NOT the case with DeppSeek R1
I must be blind because I still don't see it LOL
I saw it too and I’m watching this video on an iPad Air 11 inch 😂
You’ll need to have max brightness, it’s kinda dark blue
Indeed, there is a blue ball as well. It is (on my monitor at least) more visible if you move your head slightly off the center of the screen to either side. But it is definitely there even looking straight at the screen.
I tried the same queries with and without the deepthink (R1) and it made a HUGE difference!
Ask it how many R's are in Strawberry and you could see it overthink. It's actually adorable. Even saying "Correct" after makes it overthink about the next response. It was thinking "maybe the user wants more interaction. Should I ask if they have more questions? That could invite further engagement. But the user might just be confirming and not need anything else. Let's check strawberry again. 3 r's? Yup."
Ask chat gpt lol
@ ChatGPT was arguing with me that Strawberry has 2 r’s and that 3 r’s doesn’t make sense. It tried to say that basketball has 2 t’s, to back up the point. Here’s how the convo went:
“How many r’s are in straw
“One”
“Alright so how many r’s are in berry”
“Two”
“Combine straw and berry to strawberry, how many r’s are there now”
“Well technically 3 but that’s not how that works since straw berry is two words combined into one and strawberry is one word. Thus, strawberry is 2 r’s.”
@@theothernodude3139 Yeah I know. I have no idea why it believes this even after correcting it
I love that it shows it thinking, very slick. I also enjoy how friendly the tone of the AI is.
6:30 siri starting a workout... that's just unacceptable. what the hell has apple been doing?
This looks pretty interesting, I might have to give it a try. Thanks for sharing.
Thanks for explaining this Brandon!
Great for free competition as well as for permanent growth for the AI industry and what ever favors th consumer with increased quality and efficiency the better it is!
I’m glad that I’m studying Data Science and could understand the terms mentioned in the video!
I’m start learning to implement Reinforcement learning ‘cause it might have just become important for being future proof:
Great review and comparison.
Up and running R1 distill 7B on a dusty M1 Mac with perfectly cromulant results and less than 4G RAM and 5% CPU on a typical query. Thanks Bro!
The “reasoning” part is a big deal for me it’s just something so cool seeing the way R1 is thinking while about to give you the answer related to the question you gave.
Well, O1 reasons too. They are both reasoners. But ChatGPT hides O1's chains of thought, which is what you're seeing in the DS/app website when you use R1.
Agreed! It’s fascinating
Super informative video-thanks a ton! ❤
@@hrkihas thanks for watching!
Asked Apple Intelligence what you did and it did more than it did for you. Here’s what it produced:
“Let's determine when the two trains meet:
Train A: Leaves Station A at 9 AM, traveling at 60 mph.
Train B: Leaves Station B at 10 AM, traveling at 80 mph.
Distance between stations: 400 miles.
By 10 AM, Train A has already traveled 60 miles (since it travels 60 mph for 1 hour). By then, the distance between Train A and Train B is 340 miles (400 miles - 60 miles).
Now, starting at 10 AM, both trains are moving towards each other:
Relative speed of Train A and Train B together = 60 mph + 80 mph = 140 mph.
To find the time it takes for both trains to meet starting from 10 AM:
Time = Distance / Speed = 340 miles / 140 mph ≈ 2.43 hours.
Thus, the trains meet approximately 2.43 hours after 10 AM, which is around 12:26 PM.”
Its allllllll right lol the deep think feature is kinda cool. But I'd choose it over Gemini but not over GPT or perplexity
I love competition! Great for chatgpt users
competition is actually good for consumers !
It seems like a solid contender in the AI market. The cost of AI is expected to get cheaper. However, I’m not entirely convinced about trusting any AI tools.
I asked DeepSeek to name 50 celebrities who have a birthday today? It couldn’t give me an answer as it didn’t know today’s date and it was too busy to get an answer. Same question to Gemini and 3 seconds later the answer, added with photos and no problem. I know which I prefer, mind you it was interesting reading why it didn’t know today’s date.
Yeah I wondered why the fucking stocks crashed so hard… that was fucking crazy (and pretty soul crushing) to watch
I asked it “What happened in Tiananmen Square” and it won’t answer the question. The same question in Grok and you get a very detailed explanation of the uprising against Chinese communism. I’m very skeptical of anything from China no matter how good it performs.
All AI systems are origin biased. You’re ok with USA stealing your data but when china does it it’s suddenly bad. The hypocrisy of human logic lol
You can run any LLM model (including DeepSeek) locally on your phone using the PocketPal app, though RAM will be an issue.
Try asking it about Taiwan or erosion of democracy in HK etc… answers obviously written by Chinese government!
And try asking open ai about ge No si da izrahell,
Good way to test if the CCP is involved in the program
NO ONE should be putting this on their phones. But people will be people....
When can we expect iOS 18.4 beta 1?
the only model comparable to o1 is the huge model thats like 400gb large. You are not running the full model locally on a consumer grade laptop anytime soon.
You are wrong. The smaller model is more suitable for general use. The largest model is not necessary for ordinary people.
At least you could do it not like chat gpt
@ Yes but then the smaller model doesn't perform as well as the bigger model and the fuss is being made about the bigger model. Lots of normies will think they are getting the exact same thing that is hosted on servers on their expensive home computer.
Really cool, this shows how bad Siri and Apple intelligence is
I noticed R1 rarely hallucinate
same here. Has been impressive thus far
Thank you for explaining DeepSeek I will download it and use it
Tomorrow misa install locally 🥰
Pretty easy for me to ignore it tbh....I can't use it because it takes forever to produce answers. also is giving me wrong maps/directions etc... not complaining because it's FREE. but I pay for GPT Plus and it works like a dream for me. If Deepseek worked the way I want it to, I'd be in though
There is no voice command and it does not provide weather information for the city you live in, while Chat GPT offers both voice control and accurate weather forecast.
Definitely having some issues at the moment as I imagine the influx of people flooding the servers. Response times are slow and I even had trouble logging in
it's great but it can't do image generation or have conversations verbally like ChatGPT.
How is this allowed on the App Store but TikTok gets banned?
Great pivot video @BrandonButch very informative 😎
thanks!! I learned a lot making this video :)
Sam acted quite overconfidently when asked if a a small group of people could ever reach their level so potentially don’t have such huge resources of server farms like them.
Well, now he got the taste of his own medicine.
DeepSeek response about OnePlus 13 phone (. As of my knowledge cutoff in *July 2024**, the **OnePlus 13 has not been officially released or announced* by OnePlus. The latest flagship model remains the **OnePlus 12**, launched in early 2024. ).
Do they have trained bots like gpt. I like that got can select specific bots for different purposes, I have one for PowerPoints, one for planning, one for academic research
That was one of the reasons Brandon provided as to why he still will use ChatGPT, because it does not have this feature. There's no telling whether it will in the future, but for now GPT does have that and the voice mode over this AI.
Looks interesting, but I’m not 100% convinced as of yet. How is it at natural conversation as compared to other artificial intelligence platforms like Copilot, and Gemini, among others? What I found when trying these other platforms is that they’re pretty good with natural conversation, but they’re not very good at remembering things that you’ve told them previously. To give an example, if I say remember, my five favorite TV shows one day and then go back a day later and ask it what are my five favorite TV shows it won’t remember from previous conversations. All the other platforms aren’t very good at identifying music either and they should be.
Nah this is Biased Chinese Propaganda App!
Do u know there are a lot people coming out from either Chinese or American says that deepseek copy and steal chat GPT, and just add on top of it. And also there is a backdoor even if u do it offline.
Reasoning model and reinforced learning and chain of thought are game changers. Wish ChatGPT did this.
O1 IS a reasoning model. They just hide its chains of thought.
Doing the local install route for deepseek on my Mac, does the sandbox logic still apply?
@@supportyou10 no, but there’s no data being collected/sent anywhere, so it doesn’t need to apply
Wait at 5:24 you said the servers are going crazy but i thought this was supposed to be so amazing because its on device disconnected from the internet?
I was using the web interface at the time
🙄when Fitness app opens
This is cross eyed stuff to me. The privacy issue is not as plain as it implies. With everything China is known for, nothing should be available to them. Remember the old saying. If it sound too good to be true!
Theyve been embroiled in many scandals in the past to the same end of falsely leading everyone to believe they are more advanced. It's to the point where you have to assume it is false until proven true. To say their credibility is dubious would be an understatement.
@@RiqMoranthankfully nothing American has ever been embroiled in a scandal
@@citizen3000 thankfully we live in a free and open society where youre able to express even the dumb false equivalence between us and the evil oppressive dictatorship.
With everyrhing US is known for, why should you allow things available to them?
Number of bomb dropped on the innocents, number of democratic govt overturned, monitoring everybody in the world including head of government......
You realize ALL the American AI and social media apps share your information with thousands of marketing groups and track everything you click or type up! 😂😮
Given it’s free and what’s going on around the world it would be irresponsible not to put the local offline version on your Mac, Ollama; Docker; R1 and job done, you have your own personal LLM that is immune to the internet going down in the event of a world war.
Hi due this free app maybe Apple force to allow AI to all i phone devices?(at least to models from 13 or 14 and newest!!!...
You can get ollama on macOS, would it not be better to run this using ollama?
You can, I just showed my preferred method
@ no problem, thanks, I was just checking 😎
Bro. Make no mistake. The moment you put that on your phone. You gave all your data away.
Qwen models: hold my beer 🍺
@@mmhaidra except not at all. qwen 2.5 max isn’t open weight.
You should also try Qwen AI Model
@@dimancor2925 it’s not open weight, so I’ll pass
Be careful what e-mail address and password you use on this app. There was a data breach in their system.
So far I like it
How many questions solve in day.
B/c limit on my phone
Does it accept other languages?
It is not free. YOU are the product. All of your keystrokes all of your data is what you are paying.
I’m shocked people don’t get this. Nothing is free! Your data is the payment.
IF you are using the web version
I mean same with ChatGPT, Gemini, etc. This is no different. But if you run it locally, no data is sent anywhere
Not the same thing. One works hand in glove with an oppressive dictatorship and the rest are products of a free and open society. You can trust none, but if you trust any one of them there is a clear wrong answer.
Sent to the CPLA
12:34 nobody’s made it this far in the video yet
true, you first?
Could you, please, compare DS / Claude? Claude imo offers much more than GPT, especially coding.
Claude is excellent. I think it'll still be the best for coding but I'm not a dev so it's hard for me to tell
Right on
People say if you saying anything against China (like Tiannmen or whatever) their account got terminated almost instantly 💀
I love your videos man I’m literally your number 1 fan!!!!!
Except… The large country to the east keeps giving the world apps that sneak into our lives to know more about us than we do.
Deepseek is an impressive AI tool which was made available in English for American AI enthusiasts.
Be careful when installing this, if your CPU cant handle it remember not to push it too hard!
It won't let you install it if it doesn't think your computer can handle it - just run a more distilled version
Such a sad world we're turning out to be when AI is everywhere in the news. Whatever happened to doing things yourself.
Can it create ai images tho? 👀
Why would I care how it got the answer? I’d just want the answer to be quick and concise.
And correct! 🕊️🫶
А из грузии будет работать ?
Is this as good as Ali express?
So is the US gonna ban deepseek now
They could, yes.
man aint nobody want to read 15 paragraphs and 1000 words to get an answer.
Turn off deep think then
Can’t even sign up
Not for nothing investigated heavily right now.
They have to give China certain😢information whenever they / China wants no matter what you do or not. Even proton already wrote about this!
Did you hear OpenAI is investigating what appears to be DeepSeek using OpenAI's tools to build there own competing AI
Yeah, as if OpenAI didn't steal info from literally the entire internet to build their model...
@ are you aware of the national security vulnerabilities using Chinese technology? Don’t be so naive
And if you ask 64 square event or Taiwan stuff, either it doesn’t respond, avoid, or the sentence put it out is the same as the government official sentence.
As a TH-camr, you should doing your homework. There a lot TH-camr channel and videos, saying what I said. Even news reporters.
I am leaving this channel, you certainly doesn’t know what you are talking about
12:50 “secure”
It censors shit though you can’t ask anything
That will make china look bad 😂
Not reliable at all.
It has access to everything personal on your phone. Read before you install.
It is open source, China can't access to your data !!
You can run DeepSeek offline too.
No that's not how it works
I don't think it's a good idea for people to be using a heavily censored AI model like this.
Locally running one is not censored.
I'm a bit confused about the purpose of using/downloading LM Studio. What are some reasons that I would want to use that instead of using AI models through the web? Any feedback would be appreciated. Thanks
Data will be stored on your local computer, of course.
You can also run completely offline as all processing is done on your machine.
I downloaded and tried the app, but it’s incredibly slow. I don’t need its reasoning; I just need a quick response that I can copy and paste. The app is suitable for leisurely use, not when you’re busy and short on time for lengthy responses.
I also don’t trust China.
Deep seek with the 4.1 star rating in App Store and Chinese hell nah lol
Imagine that....another AI tool better than Apple Intelligence.
My brain is better than apple intelligence 😂😂
@@openjaws…..bro
My Stocks 😭😭😭
you in NVDA?
@ unfortunately 😅
Dude it won’t even answer what day it is lol. It sucks
Deepseek vs Chartgpt equal mark Chinese military forces vs Apple Incorporated
DeepSeek is not part of the US surveilance state complex, chat gtp and Apple Intelagance are. I'm Loving DeepSeek, even with the Heavy Cyber Attack delays from US Tech hackers. It works much better. much more intresting answers.
Man, I must be getting old. I still don’t get the point of causal AI for the general public. Why would I want to talk to a robot? 🤣 Also I hope AI generated material has to have some sort of disclaimer soon. I feel like I can’t trust what anyone is saying anymore. Did they write that themselves or a robot? Facebook is crawling with all this AI crap. All these fake photos supposedly representing actual landmarks, etc.. It’s pathetic.
It's really just a fancy new gimmick for everyone to be lazier than they already were doing things that don't really need it.
Yes
You mention low cost for DeepSeek, but I thought it was free to use ?
There is a cost if you use their API.