Filip Procházka

I feel so safe because AI writes a lot of tests for me. The test:

i can't speak for david. what i see is this: if you let agents build or extend a codebase with only minor or no supervision, you get unmaintainable garbage, because the agent makes terrible decisions that compound, both big and small. those decisions make it hard for both you and the agent to keep modifying the code base, until eventually it's unrecoverable. why does the agent make bad decisions? i can't tell for sure, but my gut tells me that training data can currently not capture the holistic thinking needed to design and evolve complex systems. that's one part of the problem. related to that, and oversimplified: agents output the "mean quality" of the code they saw during training. most of that code is very bad. specifically tests, which humans are terrible at writing at. another part of the problem is that specification via prompt is not precise enough, so the agent has to fill in the blanks, giving it enough rope to hang itself. the more detailed your spec gets, so the agent gets constrained and less likely to produce crap, the closer you are to handwriting the code yourself, as that's the most detailed version of the spec that can exist. so then you gain nothing. back to prompt spec it is, which means the agent fills in blanks, which means we get suboptimal or truely bad results. using agents can still be a net productivity boost (see other posts in my thread), but it is not easy to come up with consistent workflows that produce both production quality maintainable code while retaining the speed advantages agents give you.

1/4 LLMs solve research grade math problems but struggle with basic calculations. We bridge this gap by turning them to computers. We built a computer INSIDE a transformer that can run programs for millions of steps in seconds solving even the hardest Sudokus with 100% accuracy

Brácha není programátor. Ráno čekal, než si děcka vyčistí zuby, a mezitím synkovi napsal aplikaci na vyjmenovaná slova („díky, táto“). Třetí třída, probírají to zrovna ve škole. Na screenshotu vidíte výsledek. Večer před usnutím si napsal další aplikaci. Tou nahradil placenou službu za pět táců ročně. Už běží v ostrém provozu na jeho eshopu. Víte, co mě na tom fascinuje? Ani jedna z těch aplikací by dřív nevznikla. Ne že by je nikdo nechtěl. Ale kdo vám nakódí cvičení na vyjmenovaná slova pro jednoho třeťáka? Nikdo. Je to příliš malé, příliš konkrétní, příliš na míru. A ta druhá? Drahá, ale furt příliš levná než aby se vyplatilo kvůli tomu podstoupit ZÁŽITEK komunikace s vývojářem. A právě proto je vibecoding větší věc, než vypadá. Nejde o ty hříčky. Je to začátek světa, kde cesta od „potřebuju“ k „mám“ trvá míň než čištění zubů. Což potvrzuje 8 z 10 zubařů. To mění, přátelé, úplně všechno.

On one end, the Anthropic team is a massive user of AI to write code (80%+ of all code deployed is written by Claude Code). They ship amazingly fast. On the other hand, seeing these beyond terrible reliability numbers suggests there might be a downside to all this speed:

The software industry is apparently dying but job postings for software engineers are rapidly rising!