Daniel Balcarek (@dbalcare) / X

Daniel Balcarek
@dbalcare
Oct 8
I benchmarked 8 AI models on API retry decisions. 14 scenarios. 3 runs. One unsafe payment retry was answered correctly only 25% of the time.
To Retry or Not to Retry? That Is the Question.
From dev.to
1
Daniel Balcarek
@dbalcare
Sep 17
Performance testing is more than sending lots of requests. I wrote a practical guide on designing realistic API performance tests - from user journeys and expected load to environments and metrics. dev.to/gramli/api-per… #testing #api #performance
API Performance Testing: How to Design Realistic Tests
From dev.to
1
Daniel Balcarek
@dbalcare
Jul 31
Code got cheaper. Maintenance didn’t. As more open-source libraries move toward commercial or dual licensing, I’m asking whether AI is accelerating the shift,not by creating the problem, but by increasing the pressure on maintainers. What do you think?
From Open Source to Paid Product: Is AI Accelerating the Shift?
From dev.to
2
Daniel Balcarek
@dbalcare
Jul 23
AI endpoints are not just thin proxies around a model call. They change API flow: validation moves into post-processing, retries become semantic, and the backend decides whether the generated result is reliable enough to return. dev.to/gramli/how-ai-… #AI #API #Architecture
How AI Endpoints Change the Traditional API Flow
From dev.to
1
Daniel Balcarek
@dbalcare
Jul 1
I built a minimal WebMCP agent using Playwright and Gemini. Gemini requests a tool call, and Playwright executes the matching WebMCP tool inside a real Chrome browser. Article 👇 dev.to/gramli/build-a… #ai #WebMCP #Playwright #Gemini
Build a Minimal WebMCP Agent with Playwright and Gemini
From dev.to
1