Umer Farooq

Sharing the messy path from engineer → indie founder (

[Umer Farooq @umership]

Built a CLI-first library for me to bootstrap a lot of different projects and push MVPs out faster with production-grade designs instead of starting with a blank prompt. Also added an explore option to it so it takes a project idea and explores different design direction and I commit to one.
https://t.co/o0LBhGDAGj. given 5 percent success rate for new projects need as much help as possible

[Umer Farooq @umership]

@sflorimm Scale the server and db vertically i.e increase the tiers, and then diagnose the choking points usually it's always the I/O database etc and see if some performance optimizations can be put in place

[Umer Farooq @umership]

@UnderSecE Yeah stealing is bad.

[Umer Farooq @umership]

@thebuggeddev @Kimi_Moonshot Nicee

[Umer Farooq @umership]

@Kimi_Moonshot GOAT company.
Running Agent swarm in the kimi cli right now and loving it!

[Umer Farooq @umership]

I think these things are now counter productive on who did distillation and who took the entire codebase of their users.
There was a time where anthropic had to pay 1.5billion to settle law suits over using copywrited material. Cursor just released composer without telling anyone it was kimi under the hood until someone found out so it has been happening across the industry.

[Umer Farooq @umership]

@jakecastilloooo Is there a good platform to find creators in different niches across different plarforms?

[enzo @enzo_gte]

I've been using Kimi K3 for ~16 hours now.

The model is clearly good at a lot of different things (especially frontend), but non obvious reason why people are enjoying it so much is that it clearly does not follow the same rules in terms of safeguards and copyright.

Kimi will happily clone MacOSX. If you ask it to help you improve another AI model, it will do it with a smile on its virtual face.

Ask Fable to do the same thing? It literally starts to perceive you as a criminal committing a war crime (like no bro, all I want to do is fine tune an open source model).

After using all three recent releases, Fable, GPT 5.6, and now Kimi, it's clear that the full power of the models has been significantly held back by the safeguard restrictions caused by last months debacle with the USG -- leading to the top models being quite literally lobotomized in some areas, which leads to subpar results as the safeguards pollute its entire thinking and problem solving abilities.

The funny part? Is that you could have predicted this outcome 2-3 years ago when you started to see the rise of Chinese EVs and smartphones compared to western alternatives.

They quite literally tried to copy the Tesla Model S and iPhone as hard as possible and then eventually it started to diverge to the point where their EVs and phones are just genuinely better (which is why we have export controls banning their EVs, because they would literally drive all US manufacturers to ZERO)

There is a very clear behavior difference in Chinese capitalism and American capitalism.

American capitalism tries to protects copyright, patents, etc (oh no, you can't download a book through LibGen, that's ILLEGAL!).

Versus Chinese capitalism actually just does not give a fuck.

"Hey you want a video gen model (Seeddance 2.5) trained on every single anime ever? And you want the main character to look exactly like Messi? Sure, here you go!"

You see what I mean? When one half of the competition is being held up by regulators and restrictions on people who don't understand the technology and the other half has a leader who quite literally today said they are going to set up AI centers around the world to help other countries onboard to their open-source AIs, this is the sort of results that you will start to get.

These models were not smart enough to have this difference in philosophy matter -- but the newest class of models is where this difference makes a big deal. If these models are finally at the point where they are smarter than 99% of humans, why would you want to use the American one who tries to impose its world view onto you versus the Chinese one who will just do what you say without asking any questions?

And this isn't a full on bullpost on Kimi, the model is clearly not as smart as Fable / GPT 5.6 on things like math and science, but it's lack of handcuffs means that it can show the world what the frontier labs are gatekeeping from you and that starts to build customer resentment and loyalty towards the East, which is probably not what the USG wants.

Interesting times. Interesting times, indeed.

[Umer Farooq @umership]

@0xPaulius They won't think twice before throwing customers under the bus, started from day one with flaky policies on external harnesses. The customers might be reciprocating the same as well as soon as something better comes everyone jumps ship.

[Umer Farooq @umership]

@0xPrajwal_ Enterprise just don't want to keep chasing the next shiny thing and the startegy usually is for multiple quarters. Anthropic had first mover advantage with claude code unless something drastically gets changed not many might switch any of their stack.

[Umer Farooq @umership]

Fable was competing with chatgpt 5.6 and then kimi k3 came out of syllabus, did some UI work with it, and it's really really good, the "taste" people talk about, it has plenty of it.

[Umer Farooq @umership]

@thsottiaux 🤣🤣

[Umer Farooq @umership]

@LLMJunky Good thing I am not building rocket league and just building a normal web/mobile app , with good taste skills things should be fine.

[Umer Farooq @umership]

@nrehiew_ Now the race will have to be towards value best performance to cost ratio, my entire feed is filled with how good k3 is compared to current frontier models.

[Umer Farooq @umership]

@arena @Kimi_Moonshot Love this what an insane month July has been

[Umer Farooq @umership]

@KimiDevs Legends!

[Umer Farooq @umership]

Yeah it's pretty hard to get actual numbers, it's just your own personal experience same projects same workflows and seeing how quickly the limits are being drained. Notice it more now especially with loops where it is verifying it's own work and lots of logs and repitive tokens being ingested.

[Umer Farooq @umership]

running ponytail and the rtk claude plugin and my limits are lasting twice as long now. the changes feel deliberate, reusing code instead of rewriting the same thing over and over. have been running goals 24/7

https://t.co/OSiuuRphgE
https://t.co/A9TMXyWkkc

[Umer Farooq @umership]

trying to ship products shotgun-style, marketing locked in before a single line of code, only proven-market ideas with validated and proven MRR. fable to plan, Opus on frontend, GPT 5.6 sol on backend. Using codex image gen skills for generating images.

[Umer Farooq @umership]

Competition ftw