Yousuf Jawwad

Founder @quantmHQ | Helping software teams release 15x faster, with confidence.
Past @Verizon and @Aol

[Yousuf Jawwad @debuggerpk]

@JakeFleshner automated grading

[Yousuf Jawwad @debuggerpk]

@darrenmarble apetite (for pain and success)

[Yousuf Jawwad @debuggerpk]

@boardyai building AI that turns messy real-world inputs into workflows, like handwritten grading across 13 languages with real teachers waiting. Boardy Pro would be huge for this

[Yousuf Jawwad @debuggerpk]

Finally a product market fit.

[Yousuf Jawwad @debuggerpk]

My take.

There are a lot of parallels between frontier model companies and drug dealers.

They get you started for free, then they charge for it. Those who are training models, I feel they are optimizing for token output instead of results.

[Gergely Orosz @GergelyOrosz]

Either Sam Altman reads The Pragmatic Engineer (I wrote about this in detail exactly one week ago for paid subscribers) or I'm just a week ahead to what he also sees!

[Yousuf Jawwad @debuggerpk]

Google is raising $80 billion for AI infrastructure.

Meanwhile, in Pakistan, massive electricity costs—higher than anywhere else in the region—prevent us from being remotely eligible to even be bottom feeders

[Yousuf Jawwad @debuggerpk]

@rakyll try mercury 2.

[Yousuf Jawwad @debuggerpk]

Google used to show your search results down to the millisecond. It was their way of proving how quickly they could process the web.

With Gemini 3.5, Google is going back to those roots. Speed was the MOAT then, speed is the MOAT now.

[Yousuf Jawwad @debuggerpk]

after angular 1, the OG of frontend frameworks, @sveltejs is the only framework i really enjoy.

[Yousuf Jawwad @debuggerpk]

@rakyll Google is one company I wish to work for.

[Yousuf Jawwad @debuggerpk]

@mitchellh I have this in my prompt.

"An good engineer understands that it all a tradeoff. Please use verbiage that is less cock-sure. My PTSD gets triggered."

It works most of the time.

[Yousuf Jawwad @debuggerpk]

Gold Rush? https://t.co/b4AhjT7vW5

[Yousuf Jawwad @debuggerpk]

guardrails, guardrails, guardrails ...

[Haider. @haider1]

Google Jeff Dean says bigger context windows alone are not enough

What matters is staged retrieval: lightweight mechanisms that narrow a trillion tokens down to 10 million, then to the million you actually need

"you don't need a trillion at once, you need the right million"

[Yousuf Jawwad @debuggerpk]

@TheMaran 30 to 60kg from a tree only? very odd. A mango tree, 10/12 yaers old, can manage between 500 to 1200 kg easy.

[Yousuf Jawwad @debuggerpk]

100 page / sec ? on cpu?

[Oliver Prompts @oliviscusAI]

someone just open-sourced a tool that converts pdfs to markdown at 100 pages per second.

100% free. runs entirely on cpu. no expensive gpus needed.

[Yousuf Jawwad @debuggerpk]

Gift Indeed.

[John Carmack @ID_AA_Carmack]

I know there is some overlap between open source and anti-AI activists, but I have a hard time reconciling it. My million+ open source LOC were always intended as a gift to the world. Yes, I would make arguments about how it would strengthen our communities, and the GPL would prevent outright exploitation by our competitors, but those were to allay fears of my partners to allow me to make the gift.

AI training on the code magnifies the value of the gift. I am enthusiastic about it!

Some people do look at open source as a tool for social change, career advancement, or reputation building, but those are all downstream of the gift.

[Yousuf Jawwad @debuggerpk]

F-15 had never been shot down before this.

https://t.co/4ZlzMZLzST

[Yousuf Jawwad @debuggerpk]

another @JeffDean joke.

[Yousuf Jawwad @debuggerpk]

cloudrun ... FTW!

[Branko @brankopetric00]

New startup. You're the first infrastructure hire.

CTO wants the platform ready in 3 months.

Option A: Kubernetes from day one
- Future-proof
- Steep learning curve
- Longer initial setup
- Easier to scale later

Option B: Simple ECS + Fargate
- Faster to production
- Less operational overhead
- Might need migration later
- AWS lock-in

Option C: Just EC2 + Docker Compose
- Fastest MVP
- Manual scaling
- Technical debt guaranteed
- Cheapest short-term

Team size: 4 developers (none with K8s experience)
Expected growth: 10x users in 12 months
Funding: Series A secured

What's your recommendation?