Will Cullen
itzcull.com
Will Cullen
@itzcull.com
The data I'd love to see...

The total text "throughput" between different organisation members and channels on a communication platform like Slack.

I'd love to see if certain information flow graph patterns and structures correlate with business success
August 14, 2025 at 3:30 AM
If we assume LLM & AI capabilities will continue to improve, what skills should you invest in now, such that your positioning in the market is positively correlated to those improvements?
August 9, 2025 at 12:31 AM
Human or AI, consider the following...

1. Does the software behave as expected?
2. Can you guarantee the software behaves as expected?
3. Can the software be changed easily?
4. Does it solve your user's problem?

If you can answer "Yes" to these questions, does the "how" matter?
July 29, 2025 at 8:26 AM
Real "Agile" is making the synchronisation of product and engineering fast, stable & continuous.

Product is always ahead of Engineering, the goal is to minimise distance.
July 14, 2025 at 4:21 AM
Here's a nice workflow I've found for keeping my personal software ecosystem up to date with AI
June 19, 2025 at 10:58 PM
Can't wait for 'vibe coders' to build the medical & infrastructure software!
March 14, 2025 at 4:15 AM
So excited to see TypeScript get a native implementation

Golang is such a good choice for this, building artefacts for many different kinds of architectures and devices is an absolute breeze!

Red squiggly's are coming FASTER to an IDE near you!

visualstudiomagazine.com/Articles/20...
Microsoft Ports TypeScript to Go for 10x Native Performance Gains -- Visual Studio Magazine
Microsoft is revamping its TypeScript programming language with a native compiler and toolset. This effort seeks to address performance challenges, especially in large codebases, by porting the existing TypeScript compiler from TypeScript/JavaScript to the native language, Go.
visualstudiomagazine.com
March 12, 2025 at 12:59 AM
There's a big difference between learning with action, and learning with thought

These are our Implicit & Explicit memory systems

Explicit, means knowing "why", without knowing "how"

Implicit, means knowing "why", but not "how"

Having both is the real key to understanding
March 3, 2025 at 1:03 AM
Can someone give me a "technical" definition for what they think "Clean Code" is?
February 26, 2025 at 7:54 AM
Your codebase is like a garden.

If you do not tend to your garden, it will rot, and die.
February 21, 2025 at 12:57 AM
If you aren't matching or exceeding the standard this genius CTO has set, your software product will lose.

Linear is killing it, the UX bar has been raised.

youtu.be/XTMyOtvBJ_g
#15 – Tuomas Artman: Linear, sync engines, rethought startup MVP
The guest of this episode is Tuomas Artman, co-founder and CTO of Linear. Prior to Linear, Tuomas had already built sync engines for over a decade at compani...
www.youtube.com
February 18, 2025 at 10:04 PM
The Content Marketing Funnel boils down to 2 things...

1. Give high quality information for FREE
2. SELL putting that information into practice
February 18, 2025 at 2:32 AM
The ability to ask better questions has never been more valuable
February 15, 2025 at 10:00 PM
Learning isn't always positive

You can learn the wrong thing. You can have the wrong assumptions. You can become addicted.

Without feedback from something or someone who knows more or better than you, you'll never know if you're on the path to being less wrong

Feedback is key.
February 6, 2025 at 1:00 AM
I hear lots of DeepSeek R1 talk, but no talk about Cloudflare's R2.

Do better BlueSky.
February 4, 2025 at 10:00 PM
Can you recall what you learned from your AI of choice yesterday?

AI may help productivity, but without deliberate effort it doesn't help us understand anything any better

It might free us from needing to know more, but the cost is lack of deep understanding

No bueno
February 3, 2025 at 10:00 PM
Tech startups are missing out...

Using IaC (Infrastructure as Code) tools gives crazy operational efficiency.

Don't make the mistake of forgoing it's use, especially since you've got the luxury of being greenfield.

Pick SST or Pulumi, even Docker compose is a good start.
February 2, 2025 at 10:00 PM
No matter your software design philosophy, there are 3 things that should always be achieved in a good software solution.

1. It can serve your users reliably
2. The source code communicates intent effectively
3. It is easy to change

What else do you think matters?
January 30, 2025 at 2:23 AM
GPT Wrappers - 1
Foundational Models - 0

Your move.
January 28, 2025 at 9:01 AM
The point at which people cannot reasonably keep up with what's important in their area of expertise, is the point at which software focused on learning & retention will flourish.
January 27, 2025 at 10:22 PM
A neat little trick for improving retention when taking notes from the Internet..

Make sure your Notes app & learning resource aren't visible at the same time

You will force yourself to recall the content in your own words & models, which encodes it better within your own brain
January 27, 2025 at 6:25 AM
I love TailwindCSS, but there's a downside

The tool doesn't encourage developers to think about design tokens more semantically

The default theme is too good, and keeps them from going beyond thinking in CSS properties, and instead applying design decisions more holistically
January 25, 2025 at 7:35 AM
Customers don't care about Unit Tests, or Clean Architecture.

Nor do they matter in isolation

What they do is help us deliver software that works reliably, in a sustainable manner, while continually adapting to changing market demands with lower risk.

They keep us in the game
January 24, 2025 at 5:54 AM
The time you spend thinking about an AI's response, is inversely proportional to the likelihood it'll replace you.

More questions and more thinking is better.
January 22, 2025 at 10:00 PM
I worry that our usage of AI will take us out of the "Learning Loop".

It's fine to rely on it for mundanity, but there's still craft we can immerse ourselves in that will pay dividends for our mental evolution.

Even if it merely means better quality training data for the LLMs.
January 22, 2025 at 11:14 AM