close

DEV Community

Dimitris Kyrkos  profile picture

Dimitris Kyrkos

I'm a math grad, working as a Market Analyst for Cyclopt.

Location Thessaloniki, Greece Joined Joined on  Email address dkyrkos@cyclopt.com Personal website https://www.cyclopt.com/

Education

Aristotle University of Thessaloniki

Work

Market Analyst

Your AI-generated tests aren't testing your code. They're testing the AI's blind spots.

Commenters share how wrong algorithms still pass

Your AI-generated tests aren't testing your code. They're testing the AI's blind spots.

BERJAYA BERJAYA BERJAYA 16
Picked as gem Comments 11
3 min read

Want to connect with Dimitris Kyrkos ?

Create an account to connect with Dimitris Kyrkos . You can also sign in below to proceed if you already have an account.

Already have an account? Sign in
Semantic caching isn't a cost-saving hack. It's an admission that most "AI features" are FAQ bots in disguise.

Semantic caching isn't a cost-saving hack. It's an admission that most "AI features" are FAQ bots in disguise.

BERJAYA BERJAYA BERJAYA 20
Comments 6
3 min read
AI autocomplete isn't a productivity tool. It's a judgment test you take every few seconds.

AI autocomplete isn't a productivity tool. It's a judgment test you take every few seconds.

BERJAYA BERJAYA BERJAYA 13
Comments
3 min read
AI-Generated Code Doesn't Have a Readability Problem. It Has a Style-Enforcement Problem Wearing an AI Hat.

AI-Generated Code Doesn't Have a Readability Problem. It Has a Style-Enforcement Problem Wearing an AI Hat.

BERJAYA BERJAYA BERJAYA 22
Comments 4
3 min read
Using AI to Code Isn't the Risk. Not Understanding What It Shipped Is

Pragmatic gut checks for merging AI code

Using AI to Code Isn't the Risk. Not Understanding What It Shipped Is

BERJAYA BERJAYA BERJAYA 22
Comments 12
3 min read
Your model doesn't need to pass the bar exam. It needs to parse a log file.

Your model doesn't need to pass the bar exam. It needs to parse a log file.

BERJAYA BERJAYA BERJAYA 24
Comments 25
3 min read
The all-purpose agent isn't an architecture. It's a single point of failure with a system prompt.

The all-purpose agent isn't an architecture. It's a single point of failure with a system prompt.

BERJAYA BERJAYA BERJAYA 21
Comments 9
3 min read
"Unlimited context" is not a feature. It's technical debt with better marketing.

Production latency and token cost pitfalls

"Unlimited context" is not a feature. It's technical debt with better marketing.

BERJAYA BERJAYA BERJAYA 40
Comments 36
3 min read
Banning ChatGPT Won't Fix Your Shadow AI Problem. It'll Just Hide It.

Banning ChatGPT Won't Fix Your Shadow AI Problem. It'll Just Hide It.

BERJAYA BERJAYA BERJAYA 20
Comments 17
3 min read
RAG isn't an AI problem. It's a data engineering problem wearing an AI hat.

RAG isn't an AI problem. It's a data engineering problem wearing an AI hat.

BERJAYA BERJAYA BERJAYA 25
Comments 17
4 min read
The fallacy of "AI-first." Start with the friction, not the technology.

The fallacy of "AI-first." Start with the friction, not the technology.

BERJAYA BERJAYA BERJAYA 30
Comments 14
3 min read
AI frameworks make the first 10% feel like magic. The other 90% is where they break you.

AI frameworks make the first 10% feel like magic. The other 90% is where they break you.

BERJAYA BERJAYA 19
Comments 19
3 min read
The AI conversation is shifting from "what can it do" to "can we rely on it"

The AI conversation is shifting from "what can it do" to "can we rely on it"

BERJAYA BERJAYA BERJAYA 34
Comments 25
3 min read
The biggest barrier to enterprise AI adoption isn't the model. It's trust in everything around it.

The biggest barrier to enterprise AI adoption isn't the model. It's trust in everything around it.

BERJAYA BERJAYA 22
Comments 1
3 min read
Attackers are hijacking exposed AI endpoints to run offensive operations. No exploit needed.

Attackers are hijacking exposed AI endpoints to run offensive operations. No exploit needed.

BERJAYA 16
Comments 4
4 min read
5 Even More Advanced Java Tips That Senior Engineers Actually Use

5 Even More Advanced Java Tips That Senior Engineers Actually Use

BERJAYA 15
Comments
9 min read
Functional doesn't mean correct. That's the biggest risk with AI-generated code.

AI skipping critical requirement friction

Functional doesn't mean correct. That's the biggest risk with AI-generated code.

BERJAYA BERJAYA BERJAYA 29
Comments 67
3 min read
AI isn't a software upgrade. It's an organizational redesign.

AI isn't a software upgrade. It's an organizational redesign.

BERJAYA 17
Comments 12
2 min read
AI makes writing code easier. It doesn't make engineering easier.

The widening gap between code and context

AI makes writing code easier. It doesn't make engineering easier.

BERJAYA BERJAYA BERJAYA 29
Comments 32
2 min read
5 More Advanced Java Tips That Senior Engineers Actually Use

5 More Advanced Java Tips That Senior Engineers Actually Use

BERJAYA 8
Comments 1
9 min read
The Novo Nordisk breach hit patient data and proprietary AI models. The attack surface is expanding.

The Novo Nordisk breach hit patient data and proprietary AI models. The attack surface is expanding.

BERJAYA 10
Comments
3 min read
5 Advanced Java Tips That Senior Engineers Actually Use

5 Advanced Java Tips That Senior Engineers Actually Use

BERJAYA 9
Comments 4
6 min read
Vibe coding should accelerate development, not bypass validation

Vibe coding should accelerate development, not bypass validation

BERJAYA 8
Comments 3
2 min read
AI's real value isn't automation. It's how fast you can act on what you already know.

AI's real value isn't automation. It's how fast you can act on what you already know.

BERJAYA 5
Comments
2 min read
2026's breach list is out and the pattern is depressingly familiar

2026's breach list is out and the pattern is depressingly familiar

BERJAYA 4
Comments
2 min read
Vibe coding gets you the prototype. Then what?

Vibe coding gets you the prototype. Then what?

BERJAYA 9
Comments
2 min read
AI projects don't fail because the model is bad. They fail because nobody planned for the mess around it.

AI projects don't fail because the model is bad. They fail because nobody planned for the mess around it.

BERJAYA 5
Comments
1 min read
LLMs Generate Code, But They Can't Absorb Accountability

LLMs Generate Code, But They Can't Absorb Accountability

BERJAYA 2
Comments
4 min read
185,000 Affected in 7-Eleven Breach: Why Salesforce Is the New Soft Target for ShinyHunters

185,000 Affected in 7-Eleven Breach: Why Salesforce Is the New Soft Target for ShinyHunters

BERJAYA 1
Comments
4 min read
Most Organizations Don't Have an AI Problem, They Have an Integration Problem

Most Organizations Don't Have an AI Problem, They Have an Integration Problem

BERJAYA 8
Comments
4 min read
The GitHub Breach Is a Wake-Up Call: Your Dev Toolchain Is an Attack Surface

The GitHub Breach Is a Wake-Up Call: Your Dev Toolchain Is an Attack Surface

BERJAYA 3
Comments
3 min read
Vibe Coding Is Great for Starting, But Not for Knowing

Vibe Coding Is Great for Starting, But Not for Knowing

BERJAYA 3
Comments
5 min read
The OpenAI Breach Wasn't About OpenAI – It Was About the 84 Packages Above Them

The OpenAI Breach Wasn't About OpenAI – It Was About the 84 Packages Above Them

Comments
2 min read
Why Your AI Product Breaks in Production: It's a Distributed Systems Problem, Not a Model Problem

Why Your AI Product Breaks in Production: It's a Distributed Systems Problem, Not a Model Problem

BERJAYA 2
Comments
4 min read
Why Most AI-Generated Codebases Accumulate Invisible Technical Debt Faster Than Expected

Why Most AI-Generated Codebases Accumulate Invisible Technical Debt Faster Than Expected

BERJAYA 4
Comments 2
5 min read
State-Sponsored Hackers Are Exploiting Palo Alto Firewalls Right Now – And There's No Patch Yet

State-Sponsored Hackers Are Exploiting Palo Alto Firewalls Right Now – And There's No Patch Yet

BERJAYA 3
Comments
3 min read
2.45 Billion Requests, 1.2 Million IPs: Why Traditional Rate Limiting Is Dead

2.45 Billion Requests, 1.2 Million IPs: Why Traditional Rate Limiting Is Dead

BERJAYA 2
Comments
4 min read
When the Platform Your School Trusts Gets Hacked, Who's Actually Responsible?

When the Platform Your School Trusts Gets Hacked, Who's Actually Responsible?

BERJAYA 5
Comments
4 min read
Why Most AI Developer Tools Fail (It's Not What You Think)

Why Most AI Developer Tools Fail (It's Not What You Think)

BERJAYA 4
Comments
4 min read
AI Guardrails in Production: The Boring Engineering That Makes AI Features Actually Work

AI Guardrails in Production: The Boring Engineering That Makes AI Features Actually Work

BERJAYA 2
Comments
6 min read
"Beyond Linting: A Data-Driven Approach to Suggesting Better Code, Not Just Flagging Bad Code"

"Beyond Linting: A Data-Driven Approach to Suggesting Better Code, Not Just Flagging Bad Code"

BERJAYA BERJAYA 5
Comments 2
3 min read
Why Debugging AI-Generated Code Feels Harder Than It Should

Why Debugging AI-Generated Code Feels Harder Than It Should

BERJAYA 4
Comments 3
4 min read
Building AI Systems vs. AI Features: What Nobody Tells You About Production

Building AI Systems vs. AI Features: What Nobody Tells You About Production

BERJAYA 2
Comments
5 min read
100+ Data Breaches in Two Weeks: Why Security Can't Be an Afterthought in Your Code

100+ Data Breaches in Two Weeks: Why Security Can't Be an Afterthought in Your Code

BERJAYA BERJAYA 4
Comments 2
3 min read
Stop Writing Features, Start Building Systems: The Secret to Coding with AI

Stop Writing Features, Start Building Systems: The Secret to Coding with AI

BERJAYA 4
Comments
2 min read
Anthropic's Claude Managed Agents: 10x Speed, but at What Security Cost?

Anthropic's Claude Managed Agents: 10x Speed, but at What Security Cost?

BERJAYA 3
Comments
2 min read
The "Vibecoding" Debt Bomb: Why AI Code is Architecturally Radioactive

The "Vibecoding" Debt Bomb: Why AI Code is Architecturally Radioactive

BERJAYA 3
Comments
3 min read
The Verification Paradox: Why 100% of AI-Assisted Devs Face Incidents

The Verification Paradox: Why 100% of AI-Assisted Devs Face Incidents

BERJAYA 5
Comments 9
2 min read
5 Python Engineering Patterns for Resilience and Scale

5 Python Engineering Patterns for Resilience and Scale

BERJAYA BERJAYA 2
Comments
4 min read
5 More Advanced Python Patterns for High-Scale Engineering

5 More Advanced Python Patterns for High-Scale Engineering

BERJAYA BERJAYA 3
Comments
4 min read
The Model Collapse Paradox: Why Your 2026 AI Strategy is a House of Cards

The Model Collapse Paradox: Why Your 2026 AI Strategy is a House of Cards

BERJAYA BERJAYA BERJAYA 3
Comments 11
2 min read
The EU Security Pincer: Why You Can’t Solve NIS2 Without the Cyber Resilience Act (CRA)

The EU Security Pincer: Why You Can’t Solve NIS2 Without the Cyber Resilience Act (CRA)

BERJAYA 4
Comments
3 min read
Vibe Coding vs. Reality: Why Your AI-Generated Code Needs DevSecOps

Vibe Coding vs. Reality: Why Your AI-Generated Code Needs DevSecOps

BERJAYA 3
Comments
3 min read
5 Advanced Python Tips That Senior Engineers Actually Use

5 Advanced Python Tips That Senior Engineers Actually Use

BERJAYA 4
Comments
6 min read
The $15 Million Risk: Why "Good Enough" Code is Bankrupting Your Business

The $15 Million Risk: Why "Good Enough" Code is Bankrupting Your Business

BERJAYA 3
Comments
2 min read
The Unseen Math Behind Your SAST Tool: How Static Analysis Works Its Magic

The Unseen Math Behind Your SAST Tool: How Static Analysis Works Its Magic

BERJAYA 2
Comments
4 min read
5 Advanced JS Patterns to Level Up Your Senior Engineering Game

5 Advanced JS Patterns to Level Up Your Senior Engineering Game

BERJAYA 1
Comments 2
2 min read
Beginner Tips For A Maintainable Code

Beginner Tips For A Maintainable Code

Comments 3
3 min read
Elevate Your Code Reviews: Moving Beyond Bug Hunting

Elevate Your Code Reviews: Moving Beyond Bug Hunting

BERJAYA BERJAYA BERJAYA 5
Comments
2 min read
Monetizing Quality Assurance: Demonstrating the Financial Impact of QA Investments

Monetizing Quality Assurance: Demonstrating the Financial Impact of QA Investments

BERJAYA 1
Comments
3 min read
loading...