Viewing profile — mzubairtahir

Harness can totally change capability of the model

Openai is optimizing for gpqa diamond and anthropic is optimizing for humanity last exam. gpt 5.6 wins on gpqa and opus 5 wins on humanity last exam

I was checking benchmarks of muse spark and first it looked impressive that it scores 80 on terminal bench 2.1, but then i saw vals ai's post: in which they are showing that muse s…

Some moats are going to change in future

It is

That's true that businesses don't want to waste their time by maintaining that software but it's also true that a bigger number of users will start cloning in the future especially…

It's not about cloning and selling your own app. It's about using app that you are paying for, by cloning it using AI.

You did not get my point. My point was about if user just start cloning apps they are already using.

There is a trend that is just starting: People are cloning softwares and websites using Ai(fable 5 is doing great), what do you think what things would protect businesses from this…

i mean currently people are burn out due to a lot of ai models mainly due to decision fatigue which one to use, what will happen when we would have millions?

what? both are two different things

If we imagine future it is clear that there would be millions of ai models. then how we would navigate them?

THinking about making a open source community where developers using different models could evaluate for different tasks. Imagine small model doing similarly like large model for a…

imporvements are happening in other things like less token usage, speed, cost etc