Hey Corbin, Just wanted to say I've been really exploring a bunch of different AI tools since starting to watch your videos about a month ago. It's absolutely wild how fast this space is moving. Thanks for your videos!
I'm the director of IT & security for a company. I personally use o1 Pro all the time for cyber threat intel and security architecture R&D. The way I use it simply isn't possible with R1. However, for the average user, R1 is most definitely the best free go-to for 99% of prompts.
director role will not increase your statement about o1 Pro is much better. Director role is prove that you have made a significant contribution to your company. Benchmark make the talking, not because you are director....😅
I wouldn't trust a test that just challenges the model on a static frontend. Frontend design has rarely been ChatGPT's forte (Claude is way better). I've tried using deepseek in my app project, and while it's good, o1-pro one-shots feature requests with almost no errors. I built a PERN-stack chat app for me to use the chatgpt apis, and after Operator was released I asked o1-pro to implement the same thing. It one-shotted (a basic) version of it and added the functionality across the stack.
What was your hardware specs since you installed the model locally, and how long did it take ? - I think people are missing the point here. Hardware matters because your running it locally, why are people missing this ? I have a dedicated workstation to run this with 3 A6000's and it still manages to take time. Granted* ive have an abundance of data it's trained on. But my machine is pulling almost 1 mw, in power each month so to run DSR1 efficiently it's still using almost $150 in power alone.
do you guys believe that there will be deepseek co pilot for free extension in visual studio using r1 or r3 similar to github copilot but for free?? and how many time will it take to happen?
100%, all you need to know is the concepts of programming, coding is done with these models. I never code considering how quick these models produce incredible outputs, I use DeepSeek API inside of Cursor for VSC in building applications this way.
@@xcviij7045 thanks man, like this is exactly something I was looking for. Like I was wondering, once I know the concepts and the syntax of a language, I can go directly to coer with ai. Thanks for this
@@HiddenOne19 I understand man, assistant can still assist me in coding. Interesting times ahead so anything can happen in this dynamic world. Anthropic CEO just said ai can replace coders, so let's see how god does this
Bro Can I learn coding on your page 🌪️🌪️ Open Ai has been putting me on the bend, But now I found a group that are the God's of Ai Let's goooo... glad I found your account today, I heard deepseek is Open source and you can also customize your personal Robot 🤖 , It's crazy bro. This is next level
How To Install DeepSeek To Your Computer (in 10 mins)
th-cam.com/video/zrNKfiCuqCs/w-d-xo.html
You are one of the few AI channels I appreciate lol
Frrrr
Hey Corbin,
Just wanted to say I've been really exploring a bunch of different AI tools since starting to watch your videos about a month ago. It's absolutely wild how fast this space is moving. Thanks for your videos!
I'm the director of IT & security for a company. I personally use o1 Pro all the time for cyber threat intel and security architecture R&D. The way I use it simply isn't possible with R1.
However, for the average user, R1 is most definitely the best free go-to for 99% of prompts.
director role will not increase your statement about o1 Pro is much better. Director role is prove that you have made a significant contribution to your company. Benchmark make the talking, not because you are director....😅
@陈国伟-k2c Thanks for the input.
Yea this def makes a difference on the output.
R1 deep dive. 🙌
I wouldn't trust a test that just challenges the model on a static frontend. Frontend design has rarely been ChatGPT's forte (Claude is way better).
I've tried using deepseek in my app project, and while it's good, o1-pro one-shots feature requests with almost no errors. I built a PERN-stack chat app for me to use the chatgpt apis, and after Operator was released I asked o1-pro to implement the same thing. It one-shotted (a basic) version of it and added the functionality across the stack.
What was your hardware specs since you installed the model locally, and how long did it take ? -
I think people are missing the point here. Hardware matters because your running it locally, why are people missing this ? I have a dedicated workstation to run this with 3 A6000's and it still manages to take time. Granted* ive have an abundance of data it's trained on. But my machine is pulling almost 1 mw, in power each month so to run DSR1 efficiently it's still using almost $150 in power alone.
I have just cancelled chatGPT subscription, and singed in to a payd plan for seaart instead, with the bucks I saved!
So let's go now how to download
th-cam.com/video/zrNKfiCuqCs/w-d-xo.html
Best outcome of this is that we hear less and less of Sam Altman's vocal fry as he begs for more funding
Keep it up bro. Great content. I'm learning a lot from you.
we need adeepsea deep dive and productivity workflow with nextjs or react please
The inference battles begin, closed Source or open source Nvidia wins
Chat GPT lost its job to AI.
do you guys believe that there will be deepseek co pilot for free extension in visual studio using r1 or r3 similar to github copilot but for free?? and how many time will it take to happen?
another great post! can't wait to see more from ya. Thanks Corbin
🤝
apple being lite: ohsht😂
I've been trying for a couple days to make an account, and it seems to have crashed out... on the API side I mean.
They're experiencing DDoS attacks, and have temporarily prevented new sign-ups.
Hey as someone entering into coding, is it possible to just learn the concepts of programming and use ai to code?
100%, all you need to know is the concepts of programming, coding is done with these models. I never code considering how quick these models produce incredible outputs, I use DeepSeek API inside of Cursor for VSC in building applications this way.
AI is meant to be an assistant. It will never fix your brain or replace a proper programmer.
@@xcviij7045 100% false information, AI is not going to do everything for you, it still requires you to know exactly what you are doing.
@@xcviij7045 thanks man, like this is exactly something I was looking for. Like I was wondering, once I know the concepts and the syntax of a language, I can go directly to coer with ai. Thanks for this
@@HiddenOne19 I understand man, assistant can still assist me in coding. Interesting times ahead so anything can happen in this dynamic world. Anthropic CEO just said ai can replace coders, so let's see how god does this
Compare against Claude
Bro Can I learn coding on your page 🌪️🌪️ Open Ai has been putting me on the bend, But now I found a group that are the God's of Ai Let's goooo... glad I found your account today, I heard deepseek is Open source and you can also customize your personal Robot 🤖 , It's crazy bro. This is next level
How can you be certain that your local copy is going to be private and secure? You don't know what the system instructions are.