

Anton Firszov
1.1K posts

@antonfrv
Software Engineer | ImageSharp performance tuner | previously worked on .NET @ Microsoft | https://t.co/WvLdpXRoYP













Prime Minister Viktor Orbán has just called to congratulate us on our victory.

Hungary, in the end, chose Europe, rejected Russia, and rejected sprawling corruption. With record turnout (78% of those eligible voting), Tisza gained supermajority. Orbán’s autocratic regime, built out, step by step, over 16 years, will promptly be dismantled. Historic!






Orbanists are becoming embarrassingly desperate. @CatherineBelton was slapped with a lawsuit by Roman Abramovich a Russian oligarch, close Putin associate, and target of Western sanctions. This is a a politically motivated smear - and a clear case of crude transnational repression.



A small setback for Islamoleftism








I see a lot of "I never used to write tests but not I just tell LLM to generate a full suit". But isn't this wrong? Thinking from first principles, if tests are definitive proof that your code works then you should put a lot of manual effort into designing good test suit, and then just let LLM oneshot the actual code to satisfy the tests. Why the common practice is the other way around? Is seems that in theory TDD should thrive in agentic-first coding, but it not. Why?
