@EvansKabubu_init
Thank you for the simplified approach to explaining the topic. I subscribed immediately as your approach is different from the rest, you don't scheme over the simple yet critical stuff. Keep producing.
@DePistolero
I found cursor ai to be unbeatable when it comes to understanding a fullstack large projects and late development... I think all models can make simple app like todo lists, websites with some data, static websites etc... but when it comes to databases, actual complex and non standard logic of the app, custom solutions, etc.. can work on the first try but modifying it, refactoring it, extending it becomes a pure nightmare... just the same as if you do it yourself :)
@TechMadDad
At 12:08 you changed the model back from Gemini 2.5 to Claude 3.5. May I ask why? Considering Gemini 2.5 is the one to beat currently and is more economical , wouldn’t it make sense to use it instead of Claude 3.5 or 3.7?
@skverskk
Nice!
@Ransuari-l1g
Pure gold, is there a chance you can do a followup video on local models alternatives? is it possible to run Gemma3/Qwen Coder with VSCode Copilot Agent?
@DavidMelechYisrael
Rate limited: are there any other alternatives that have no rate limits or token limits? Any Chinese versions?
@CodeWithNathan97
NEW GPT-4.1 The BEST Coding Model! Can it beat Gemini 2.5 PRO? GPT-4.1 Overview and Test: https://youtu.be/yTpvmOK4y4U
@sedovolosyi
I like GitHub Copilot.
@Tabproxy
Hello, sir. Do you take co-operation?
@TurkicPoseidon
Nathan is the gemini Api free forever and also is it strong ?
@TurkicPoseidon
Should i pay for Github Copilot or other tools ⚠️? Can an experienced developer give me tips ?
@florence_fernandez
Why don’t you let it run in 4:20 and fail, and let it find solution on its own, e.g. using web search tool to search for a solution? That’s a more agentic way than you must have the knowledge to solve the issue and feed to it
@rhb30001
there is no code option on top left though
@joecater894
My experience is that it can be terrible... techs not ready for it yet. Whats probably needed is a browser plugin that communicates with copilot.. that might improve it... maybe.. it might allow copilot to see front end and test, plus see console etc... without it, cannot really have a reliable agent mode. Otherwise agent mode goes galloping off a cliff... doing tonnes of dodgy edits that just break everything. Not good yet.. Another good thing would be collaboration between different AI behind the scenes to cross check edits.. that could massively increase reliability... and catch dumb AI coding decisions before they get edited. The other problem with it is its incredibly slow.. for some quite simple coding work..
@rohannikumbh8056
which browser do you use ??
@truliapro7112
Can I use ollama model?
@huutuanvo-l4b
HI NATHAN, HOW CAN I CONTACT YOU?
@KodeGenZ
Is this free like trae ai?
@CodeWithNathan97