I'm glad you liked it - thanks for letting me know. I'd like to write up a whole bunch of stuff that is less about "Kilo" and more just "this is how AI is going to work in a large team with an existing codebase"
The one that holding on my c board to jump in is that code that was produced at office were proprietary and very sensitive that so much it could produce a law suit to the company..this is the case that all ai tool never discussed on how to mitigate the only thing that we go is hosted our self with monetary limitation on the stake
Do you have your own AI inference inside the network that your team trusts? Some tools - Kilo Code among them - allow you to connect to local or "private cloud" models that can solve for that issue.
Yikes, the part where AI is compared to a junior is just .. wrong. There's no comparison between a human that's actually intelligent and a next token predictor that can go fast but not always in the right direction. I get the point your selling your AI slop generator but you don't need to go so low as making apple to orange comparisons.
> The difference is this junior developer works at 50-120 tokens per second and never needs coffee breaks.
And never learns or adapts to environment changes.
The sad part is that based on the number of emojis and "—" em dashes, this was clearly written by AI. What's the point of even commenting on socials nowdays!?
I was not expecting to like that article so much. Spot on!
I'm glad you liked it - thanks for letting me know. I'd like to write up a whole bunch of stuff that is less about "Kilo" and more just "this is how AI is going to work in a large team with an existing codebase"
Great article!
I confess I was expecting a :" And did you know that the new kilo code enterprise ships most of that?!"
LoL
Anyways, really great, I will use some of this reasoning in my next team meeting.
Thanks for sharing!
The one that holding on my c board to jump in is that code that was produced at office were proprietary and very sensitive that so much it could produce a law suit to the company..this is the case that all ai tool never discussed on how to mitigate the only thing that we go is hosted our self with monetary limitation on the stake
Do you have your own AI inference inside the network that your team trusts? Some tools - Kilo Code among them - allow you to connect to local or "private cloud" models that can solve for that issue.
The c board were ok to use the local ai but prob is they don't want to invest to buy the proper GPU n cpu...
So currently I use kilocode and free ai such grok to code tools that was not part of the company product at least to make my job easier..
Yikes, the part where AI is compared to a junior is just .. wrong. There's no comparison between a human that's actually intelligent and a next token predictor that can go fast but not always in the right direction. I get the point your selling your AI slop generator but you don't need to go so low as making apple to orange comparisons.
> The difference is this junior developer works at 50-120 tokens per second and never needs coffee breaks.
And never learns or adapts to environment changes.
The sad part is that based on the number of emojis and "—" em dashes, this was clearly written by AI. What's the point of even commenting on socials nowdays!?
...I added the emojis myself 😿