Mohamed Shams

working on source code transformation

ex SDM/SDE @msft and @amzn
ex cofounder @sahl

[Mohamed Shams @moshamsx]

@davidfowl Haha LLMs only need an http client and an api token to wreak havoc

[Mohamed Shams @moshamsx]

@b4dgerr I generally still own software design

[Mohamed Shams @moshamsx]

@vboykis That’s how I think about models too. Will definitely give Gemini a shot

[Mohamed Shams @moshamsx]

@b4dgerr I’d probably generate a landing page using AI or write one using react if I have to

[Mohamed Shams @moshamsx]

@b4dgerr You can get away with using raw js

[Mohamed Shams @moshamsx]

@rodri_shema I would not shy away from running ads, even early on. I’d want to define a concrete action for these ads (eg, engagement, leads, landing page visits, app installs etc)

[Mohamed Shams @moshamsx]

@vboykis Great stuff

[Mohamed Shams @moshamsx]

@AhmedGhazey شكرا يا أحمد على الكلمات الإيجابية وعلى الإشارة

[Mohamed Shams @moshamsx]

I think one of the most useful questions a founder can ask is:

What would convince me that I’m wrong?

It’s easy to design experiments that validate your assumptions.

It’s much harder, and often more valuable, to design ones that might invalidate them.

[Mohamed Shams @moshamsx]

@ZackKorman There is a market for unpatched zero days. When the adversary has a lot of resources, it’s better to have a realistic assessment of what is possible

[Mohamed Shams @moshamsx]

@davidfowl I think agent reviews should be done before human reviews

[Mohamed Shams @moshamsx]

@rgbdev مش كل الناس بيعرفوا يتعاملوا مع النقد والأخطاء كويس، خصوصا لو عاشوا كتير في بيئات سامة

[Mohamed Shams @moshamsx]

@CFDevelop That’s one way to see it

I see it as having system that allows a reasonably good model/human to be productive without burning tokens/attention

It also makes me capable of reviewing the code and rejecting crap quickly because I understand things deeply

[Mohamed Shams @moshamsx]

Founders are often told to hustle, move fast, and do the work.

But what exactly is the work when you don’t yet know what will succeed?

I think risk is one of the most underappreciated aspects of starting something new, particularly as a solo founder.

Consider how differently a founder and an employee experience the world.

An employee typically has a schedule, reasonably well-defined problems, colleagues to lean on, and a paycheck that reliably arrives at the end of the month.

An early-stage founder might have none of those things.

And that kind of uncertainty does interesting things to human psychology.

At times, it can make you curious, adventurous, and unusually open to experimentation. At others, the very same uncertainty can produce existential dread about the work that lies ahead.

These aren’t contradictory responses. They’re two sides of the same experience.

I’ve come to think that one of the first risks a founder needs to learn to manage is their own psychology.

Uncertainty doesn’t just affect how you feel about your work. It can change how you think about the work itself.

You might pursue an idea because you’re excited about its possibilities, overlooking evidence that it isn’t working.

Or you might abandon a promising idea because the uncertainty becomes uncomfortable, rather than because the evidence tells you to.

Neither curiosity nor fear is a substitute for judgment.

There’s plenty of conventional advice about maintaining good habits: exercise, eat well, get some sunlight, maintain a social life, and sleep regularly.

All good advice.

But these habits help maintain your capacity to deal with uncertainty. They don’t eliminate the uncertainty itself.

And perhaps that’s an important distinction.

A founder needs to learn not only how to tolerate uncertainty, but also how to reason about it without allowing their emotional responses to dictate their decisions.

Some risks can be eliminated. Others can be reduced through experimentation. Some simply have to be accepted.

Learning to distinguish between them is difficult enough without confusing your own discomfort with evidence that something is wrong.

Perhaps that’s one of the harder parts of entrepreneurship: preserving the curiosity that brought you into it while developing the discipline to reason clearly about uncertainty.

How do you remain adventurous without becoming reckless? How do you acknowledge risk without allowing it to consume your attention?

These are questions I’ve spent some time thinking about, and I intend to explore them further in subsequent posts, sharing observations and ideas as I go.

[Mohamed Shams @moshamsx]

The recent PHP vulnerabilities made me reflect on the state of some open source projects that power the internet

Open source gives everyone the opportunity to scrutinize the code

But I’ve seen projects so poorly written and tested that scrutiny alone is nowhere near sufficient

[Mohamed Shams @moshamsx]

I’ve seen a few folks argue that specs and E2E tests are all you need in the age of AI. They call it spec-driven development

I’m finding fast unit tests incredibly valuable

Getting feedback in a few seconds that your latest change works is hard to beat