Serkan Özcan

Latest posts

#Apple · #MacBook Pro · #Steve Jobs · #touchscreen · #OLED · #Technology

Jobs said "It doesn't work": now Apple is selling a touchscreen MacBook Pro

October 20, 2010. At Apple's "Back to the Mac" event, the company is explaining which ideas Mac is borrowing from iOS. The question is already hanging in the air: if the iPhone and iPad are used with a finger, why can't…

Read
#PHP · #Cybersecurity · #Open Source · #Artificial Intelligence · #System Administration

What the PHP crisis taught me: Security is a shared responsibility

The recent PHP vulnerabilities should make everyone who develops software or runs servers reconsider their responsibilities. When we put a working application online, we also take responsibility for maintaining it and…

Read
#PHP · #Cybersecurity · #PHP2Shell · #Security updates

PHP2Shell: Disclosure is not when the risk begins

PHP2Shell refers to CVE-2026-109618 and CVE-2026-109619, two critical vulnerabilities that can allow unauthenticated remote code execution in affected PHP installations. Security updates were released on October 21,…

Read
#Cybersecurity · #PHP · #Security advisory · #System administration

Türkiye’s official PHP warning: A practical response for organizations

The Turkish Cybersecurity Presidency’s PHP security bulletin TR-26-1372, dated 21 October 2026, calls on organizations to review their PHP systems. My previous post covered the technical vulnerabilities; here I want to…

Read
#PHP · #Cybersecurity · #Security Updates · #Technical Analysis

PHP’s critical security alarm: Two flaws and the reality of patching

The PHP security alarm in the October 22, 2026 news cycle concerns the language runtime beneath websites, rather than a defect limited to one application or extension. CVE-2026-109618 and CVE-2026-109619 can allow…

Read
#Artificial intelligence · #Muse Spark Open · #Local AI

Muse Spark Open: open weights, local inference, and real costs

Downloading a model and running it in your own environment creates another area of control in AI products. It can give you more direct influence over data flows, runtime behavior, and integrations. Accessible weights,…

Read
#Artificial intelligence · #Anthropic · #Independent evaluation

Internal access for AI oversight: Anthropic's next step

Testing an AI system through external questions alone may not reveal all the risks visible inside its developer. How the model was developed, how safeguards operate, and how previous incidents were handled can all affect…

Read
#Artificial intelligence · #Materials science · #Science

When AI reads crystals: a three-dimensional view of materials discovery

Knowing a material's chemical name is different from understanding how its atoms are arranged. The same components can behave differently in different structures. Materials research therefore contains an important layer…

Read
#Artificial intelligence · #Anthropic · #Technology economics

Anthropic's public benefit fund: sharing the gains from growth

As AI companies grow, who receives the economic value they create becomes a more concrete question. I find the public benefit fund announced by Anthropic on October 18, 2026 interesting because it aims to link its…

Read
#Artificial intelligence · #Child safety · #Digital privacy

Child safety in AI apps: an age label is not enough

An app being available in a store or carrying an age-appropriate rating does not establish that all its features are safe for children. AI image editors make that gap more visible: a seemingly ordinary editing interface…

Read

Something to discover

3 articles
Artificial intelligence

Creating with AI: more than a tool

Before finding the right answer, ask the right question. A few thoughts on thinking alongside AI.

Read
Automation

Small automations, more room to breathe

We do not have to automate everything. Sometimes removing a single repetition is enough.

Read
Software

Making a little space for yourself online

Thoughts that keep growing, rather than finished ideas. Why is a personal website still a lovely idea?

Read
Artificial intelligence · 9 Oct 2026 · 2 min read

Creating with AI: more than a tool

The first version of an idea is usually messy. A few sentences, an uncertain goal, and the feeling of not knowing where to begin. One of the interesting things about AI emerges in that uncertainty: it offers a workspace that can respond to our thoughts.

Context before answers

“Do this for me” can be a starting point. But the outcome changes when we explain what we want to do, who we are doing it for, and the constraints we are working within. A good prompt is often a well-defined problem.

When designing a feature, it helps to start with three questions:

  • What problem are we actually solving?
  • How will we know it works?
  • Which part could we leave out entirely?

The last question is particularly useful. As creating things becomes easier, deciding what not to create matters more.

Small loops, clear feedback

Instead of expecting a perfect result in one attempt, it is more useful to work in small pieces we can check. First a draft, then a real example, then a test. Each step makes the next one easier to understand.

Speed may come from shorter feedback loops before it comes from producing more output.

This approach applies beyond code. Outlining an article, comparing different aspects of an interface, or trying to understand a complex topic can benefit from the same loop.

The final decision is ours

A fluent answer is not necessarily a correct answer. Checking sources, running the code, and seeing the result under real conditions are still part of the work. Especially when personal data or important decisions are involved, we need to think clearly about what we share and what we trust.

Perhaps the most productive relationship is to see AI as a colleague we can use to test our thinking, rather than a box to which we hand over responsibility.

Loading…